Method for monitoring a circuit, control device and circuit

Abstract

The invention relates to a method for monitoring a circuit consisting of several power transistors connected in parallel, using the values ​​of one or more parameters from a parameter set of the circuit, wherein the respective values ​​of the parameters are determined at successive time points and an error is detected by means of a deviation between values ​​that is greater than a threshold value. The invention further relates to a control device for carrying out such a method and to a circuit with such a control device.

Description

[0001] The invention relates to a method for monitoring a circuit, an associated control device and an associated circuit.

[0002] Circuits can be used, for example, to control currents, such as selectively supplying current to specific phases of an electric motor to operate it as desired. Electric motors and other devices can draw high currents, requiring the parallel connection of multiple power transistors for switching, as a single transistor cannot handle the required current. Monitoring such power transistors is often challenging, as separate monitoring of each transistor is technically feasible but complex.

[0003] The publication DE 10 2015 206 531 A1 describes a parallel circuit of power semiconductors, each connected to a fuse. The fuses trip when a current exceeds the tripping limit.

[0004] The publication DE 10 2023 202 465 A1 describes a parallel circuit of power semiconductors, whereby the gate and source / drain voltages are compared to determine whether the control signal is being converted or whether there is a fault.

[0005] The publication DE 10 2023 202 464 A1 describes a parallel circuit of power semiconductors, whereby a fault is detected when the voltage across a gate resistor exceeds a threshold.

[0006] It is therefore an object of the invention to provide a method for monitoring a circuit consisting of several power transistors, which, compared to known designs, is implemented in an alternative manner, for example, with less effort. It is further an object of the invention to provide a control device for carrying out such a method. It is also an object of the invention to provide a circuit with power transistors and such a control device.

[0007] According to the invention, this is achieved by a method, a control device, and a circuit as defined in the respective main claims. Advantageous embodiments can be found, for example, in the respective dependent claims.

[0008] The invention relates to a method for monitoring a circuit consisting of several power transistors. The power transistors are advantageously connected in parallel. Monitoring is typically carried out using values ​​of one or more parameters from a parameter set of the circuit. The method comprises the following steps: - Determine, at successive time points, the respective value of each parameter of the parameter set, and - Determine, for one or more parameters, a deviation between a value of the parameter at a given time and a value of the parameter at a previous time, and - if the deviation is greater than a threshold, an error message will be displayed.

[0009] This method allows for the simple monitoring of power transistors without requiring separate monitoring of each individual transistor. Instead, it utilizes parameters that are expected to change only by a specific amount or relative value within a given time interval, assuming normal operation of the power transistors. These parameters might be due to temperature changes, for example. If a deviation exceeds a threshold value—that is, if it is greater than expected under normal circumstances—a malfunction can be assumed.

[0010] Power transistors typically switch a current together, which can be used, for example, to power an electric motor. By connecting several power transistors in parallel, such as two, three, four, or more, a higher current can be switched than with just one or fewer. A power transistor is typically understood to be a semiconductor device capable of switching a current on and off. A power transistor may have only two switching states, i.e., on and off. It may also be designed to be adjustable, meaning it has multiple switching states with different current conductivities or a continuously adjustable current conductivity, allowing for the setting of different currents.

[0011] The parameters are typically quantities used to evaluate the function of power transistors. Values ​​are the results of measurements or calculations, or they derive from specifications regarding the respective parameter.

[0012] The aforementioned time points typically follow one another, particularly at predetermined intervals. Specifically, the same time interval can be used between immediately consecutive time points while the procedure is running. At each time point, a value is typically measured for all parameters, or at least for one or more parameters of the parameter set. The deviation is then determined in comparison to a value measured at a previous time point. This previous time point can be immediately preceding the previous one. This can mean, in particular, that there is no other time point between the two considered time points at which a value or values ​​of parameters were determined. If this deviation is greater than the threshold value, a malfunction can be assumed. In this case, an error message is typically issued.An error message can be, for example, an internal signal in an electronic control device, but it can also be a visual, audible, and / or haptic warning to a user, such as a driver of a motor vehicle. The error message can also be stored in an error data memory. This can improve traceability.

[0013] The parameter set can, in particular, include a parameter for an offset of a characteristic curve, which can specify, in particular, the voltage across the power transistors as a function of a total current through the power transistors. Such an offset is typically the value of the characteristic curve that the voltage assumes when the total current is zero. The characteristic curve can be measured at a specific point in time and / or it can be continuously updated, especially whenever a new current-voltage pair is available. Alternatively, a single current-voltage pair can be measured to update an existing characteristic curve.

[0014] According to one interpretation, the parameter set can include as a parameter the slope of a characteristic curve, which specifies the voltage across the power transistors as a function of the total current through the power transistors. Such a slope can, in particular, provide an indication of resistance. The points already mentioned above regarding the characteristic curve also apply.

[0015] In particular, an initial value of at least one parameter, especially the slope of a characteristic curve, can be determined based on a matrix containing a voltage across the power transistors and a corresponding total current through the power transistors. This can be, in particular, a characteristic curve as mentioned above. Such a matrix can, in particular, represent the characteristic curve already mentioned. It can, for example, be designed as a two-dimensional matrix. Values ​​for the slope and / or the offset can thus be determined from the matrix. An initial value can, in particular, be understood as the first value of a parameter immediately after commissioning.

[0016] In particular, the matrix can be obtained when the circuit is started up by reading a stored matrix. This stored matrix may have been saved, for example, during a previous operation of the circuit. It may be located, for instance, in persistent or non-volatile memory. Starting up the circuit could involve, for example, switching it on for the purpose of using a load such as an electric motor. This can occur, for example, when the ignition of a motor vehicle is activated. By reading a stored matrix, the circuit monitoring system is provided with a functioning matrix, allowing changes to be tracked.

[0017] In particular, a pre-defined, stored matrix can be read in during the initial commissioning of the circuit. Initial commissioning can be, in particular, commissioning by a user that occurs for the first time after the manufacture of the circuit or a higher-level unit such as a motor vehicle. For example, this can be the case when a motor vehicle with an electric motor, powered by a circuit monitored according to the method disclosed herein, is put into operation for the first time. During subsequent commissioning of the circuit, a matrix stored during a previous operation of the circuit can be read in. Subsequent commissioning is generally understood to mean any commissioning after initial commissioning.Advantageously, a previously stored matrix is ​​used, so that it can already be updated compared to the stored matrix specified when the circuit was manufactured.

[0018] In particular, a matrix to be read can contain temperature-dependent entries. Before the parameter values ​​are determined for the first time after the circuit is commissioned, a temperature can be measured. Entries can then be used based on the measured temperature. This allows the temperature dependence of parameters to be taken into account. The matrix can therefore contain entries, where each entry is assigned to a specific temperature. In other words, the matrix can have an additional dimension representing the temperature. Both modifying and reading from the matrix can thus be temperature-dependent. After commissioning the circuit, a temperature can be measured. It can be assumed that this temperature typically prevails not only at the location of the temperature sensor but also at the location of the power transistors used in the circuit.This can be assumed particularly after extended periods of inactivity. The selection of matrix entries can then be based on this temperature.

[0019] In particular, it can be provided that, before the initial determination of the parameter values ​​after commissioning the circuit, a voltage across the power transistors and a corresponding total current through the power transistors are measured and written to the matrix. This allows the matrix to be updated. This can be done immediately after commissioning the circuit. No significant temperature influences are expected at this stage.

[0020] In particular, it can be provided that, immediately before determining the parameter values ​​at a given time, a voltage across the power transistors and a corresponding total current through the power transistors are measured and written to the matrix. This allows the matrix to be updated with such a pair of values. The parameter values ​​can then be determined using this updated matrix.

[0021] In particular, it can be provided that the matrix is ​​saved immediately after a deviation is initially detected following commissioning of the circuit. Such a saved matrix can then be read in again during a subsequent commissioning, as described above. It provides a good starting point for further monitoring.

[0022] In particular, the matrix can be stored in a neural network. A neural network can be understood as an electronic arrangement in which certain connections are implemented using software and / or hardware, such that inputs to the neural network lead to specific outputs. Using a neural network for the matrix has proven particularly advantageous. Writing values ​​to the matrix can be achieved, in particular, by making changes within the neural network.

[0023] In particular, it can be stipulated that the circuit is only considered commissioned if it has been out of operation for at least a predetermined period immediately beforehand. This ensures that commissioning is only assumed when it can be assumed that the temperature across the components, especially the power transistors and any temperature sensor that may be used, has equilibrated. In this case, the monitoring method described herein leads to particularly advantageous results.It can be assumed, in particular, that initially a measured temperature corresponds to the temperature of the power transistors. However, during operation of the circuit, it can typically be assumed that a temperature measured by a temperature sensor will deviate from the temperature of the power transistors, since heat transfer only occurs with a certain time lag. If the period during which the circuit was out of operation was too short, it can still be operated, but preferably not with the consequences described herein regarding commissioning.

[0024] In particular, it may be possible to determine only an initial value of a parameter, especially the slope of the characteristic curve, based on the matrix. In particular, an initial value of another parameter, especially the offset of the characteristic curve, may be calculated based on the initial value determined by the matrix and the characteristic curve itself. In this case, only one parameter, especially the slope, is determined by the matrix, and the other parameter, especially the offset, is calculated. The reverse of the above is also possible; that is, the offset can also be determined by the matrix, and the slope can be determined by the characteristic curve.

[0025] In particular, it may be provided that the values ​​of the parameters are determined, at least after an initial determination, by the following steps: - Measuring a voltage across the power transistors and a corresponding total current through the power transistors, - Calculating a new value of a first parameter, or an offset of the characteristic curve, based on the measured voltage and the measured associated total current, and - Calculating a value of a second parameter, or a slope of the characteristic curve, based on the measured voltage and the measured associated total current, as well as the previous value of the first parameter.

[0026] The steps mentioned above can be repeated continuously. During such a repetition or iteration, a comparison of two parameters can also be performed to detect errors.

[0027] This approach allows values ​​to be continuously updated during the execution of the procedure, particularly after initial values ​​have been determined, especially as described above. Preferably, no matrix is ​​used. Specifically, when a new measured current and voltage pair is available, a new offset value can first be calculated while maintaining the slope. This newly calculated offset can then be compared with a previous offset to detect any errors. Subsequently, a new slope can be calculated using the newly measured current and voltage pair and the previous offset value. The new slope determined in this way is then typically used when the aforementioned steps are repeated.However, it is also possible to carry out the aforementioned procedure in such a way that current and voltage are reversed compared to the version just described.

[0028] In particular, it can be provided that the steps to be executed during operation of the circuit are repeated multiple times. This ensures continuous monitoring. Specifically, the steps can be executed for as long as the circuit is in operation.

[0029] According to one implementation, the deviation for one, several, or all parameters can be determined as an absolute deviation. This allows, in particular, a comparison of a value with a previous value, calculating a difference, and this difference is then compared to the threshold value, independent of the absolute values. Alternatively, according to another implementation, the deviation for one, several, or all parameters can be determined as a relative deviation. Here, too, a difference can be calculated as described above, but this difference can then be related to an absolute value of one of the parameters, for example, by dividing the difference by the absolute value of one of the parameters. The resulting value can then be compared to the threshold value.

[0030] In particular, the power transistors may be MOSFETs (metal-oxide-semiconductor field-effect transistors). The procedure described herein has proven especially advantageous for such power transistors. However, the method can also be applied to other power transistors.

[0031] The invention further relates to a control device. This device is configured to execute a method as described herein. The invention further relates to a non-volatile, computer-readable storage medium on which program code is stored, the execution of which causes a processor to perform a method described herein. All embodiments and variants described herein may be used with regard to the method.

[0032] The invention further relates to a circuit. The circuit comprises several power transistors. The circuit includes a control device as described herein. The advantages described herein can be achieved by means of such a circuit. With regard to the control device and, in particular, with regard to the method implemented therein, all embodiments and variants described herein can be used.

[0033] In other words, the following solutions are known for diagnosing MOSFETs connected in parallel. Firstly, the individual current of each MOSFET can be measured and used for diagnosis. For example, if the current of one MOSFET increases sharply, a fault may be present. Alternatively, the total voltage of the parallel-connected MOSFETs, as well as the total current and temperature, can be measured. From this, a total resistance can be calculated. If this value deviates from a calculated value, a fault may be present. However, these measures are typically complex, especially when measuring separately, and usually only work reliably with a small number of MOSFETs connected in parallel.

[0034] As a solution, for example, a pre-trained neural network can be calibrated at an end-of-line position during the initial setup for the hardware temperature and for several current and voltage values. This neural network can have input data such as voltage, current, and temperature. It can then output, for example, a resistance value. Parameters such as the slope and / or offsets of a characteristic curve, such as U = a * I + U0, can also be calibrated at the end of the line or during production. When MOSFETs are first powered on, it can then be assumed that the temperature of the MOSFETs and the temperature sensor corresponds to that of the surrounding environment. A standby period can be used as a reference for this. During the initial power-up, the measured values ​​of voltage (U), current (I), and temperature (T) can be fed into a matrix.The resistance calculated from the matrix can then be compared with the ratio of the measured values ​​U / I. If the difference is less than zero, a copy of the matrix can be created. The difference and the measured input values ​​can be used to update the neurons of the matrix. After the new matrix is ​​created, it is used as a reference during a restart. If a difference becomes too large, an anomaly can be reported. At each restart, a resistance value can be calculated using the matrices. The difference can be used as an indication of aging of the MOSFETs used. During operation, a validated resistance value may be available, particularly after a restart. This value may be validated, meaning no anomaly is present, and can be determined from a U / I measurement. This measurement can be used to update the values ​​of the matrix or the characteristic curve.From now on, current and voltage can be measured cyclically. If the calculated differences between measured or calculated values ​​are too large, especially for voltages or resistances, an anomaly can be detected or an error message issued. Otherwise, newly measured values ​​can be used for the next cyclic calculation.

[0035] Further features and advantages will be evident to those skilled in the art from the exemplary embodiment described below with reference to the accompanying drawing. These show: Fig. 1: a circuit, Fig. 2: a matrix, and Fig. 3: a flowchart.

[0036] Fig. Figure 1 shows a circuit 10 according to an embodiment of the invention. The circuit 10 comprises a voltage source 20, which supplies a voltage to be switched. The circuit 10 also comprises a switchable voltage source 25, which performs control functions that will be discussed in more detail below. The voltage sources each have a positive terminal (+) and a negative terminal (-). For switching, the circuit 10 comprises a total of five power transistors 31, 32, 33, 34, 35, each of which is connected to the positive terminal of the voltage source 20 via one power terminal and is connected together at the bottom via the opposite power terminal. The power transistors 31, 32, 33, 34, 35 are thus connected in parallel.Their respective gate connections, by means of which the power transistors 31, 32, 33, 34, 35 can be controlled, are connected to the further voltage source 25 via respective resistors 41, 42, 43, 44, 45.

[0037] The power transistors 31, 32, 33, 34, and 35 are connected to a shunt resistor 51, across which an ammeter 50 is connected. As shown, the ammeter 50 is implemented as an operational amplifier, which measures the voltage drop across the shunt resistor 51 and thus allows the current flowing through it to be determined. Furthermore, a voltmeter 55, also implemented as an operational amplifier, is connected in parallel to the power transistors 31, 32, 33, 34, and 35 and measures the voltage drop across them. A temperature sensor 60 is provided to monitor the temperature of the circuit 10.

[0038] A control device 70 according to an embodiment of the invention is provided for monitoring the circuit 10. This device is configured to execute a method according to the invention. Such a method is described in particular below with reference to Fig. 3. The control device 70 is connected to the ammeter 50, the voltmeter 55, and the temperature sensor 60. This allows current, voltage, and temperature to be measured. Furthermore, the control device 70 is connected to the additional voltage source 25, so that the additional voltage source 25 can be switched on and off, which ultimately also allows the power transistors 31, 32, 33, 34, and 35 to be switched on and off.

[0039] Fig. Figure 2 shows a matrix Mx purely schematically. The matrix Mx is typically organized in the form of a neural network. This neural network can be trained by initial training and further training with respective combinations of voltage U, current I, and temperature T. These values ​​can also be used as input variables, and the slope a of a characteristic curve is output. Such matrices Mx are used within the framework of the now discussed Fig. The method described in section 3 is used. The characteristic curve indicates the relationship between current I and voltage U at a specific temperature T.

[0040] Fig.Figure 3 shows a procedure for monitoring circuit 10. The procedure begins with the "Start" block when the circuit is switched on or commissioned. It then checks whether circuit 10 has started after a standby period. Adherence to such a standby period typically means that circuit 10 has not been in operation for at least a predetermined time. In this case, it can be assumed that the temperature is uniform throughout the circuit and that temperature sensor 60 measures a temperature that is also present at power transistors 31, 32, 33, 34, and 35.

[0041] If commissioning occurs after a standby period, the system checks whether it is a first-time start. A first-time start is typically the very first start since the circuit was manufactured, for example, after circuit 10 has been delivered to a customer and installed in their system. If this is the case, a previously saved matrix M0 is loaded as the matrix Mx to be used. Matrix M0 can be saved during the production of circuit 10. Otherwise, a previously saved matrix M1 is loaded as the matrix Mx to be used.

[0042] The system then waits until switches are turned on. These switches are typically the power transistors 31, 32, 33, 34, and 35. As long as none of these line transistors 31, 32, 33, 34, and 35 have been turned on, the system waits for this to happen. Once a switch has been turned on, voltage U, current I, and temperature T are measured. Voltage U and current I are then fed into the matrix Mx, which has been loaded above, as input values ​​so that it can be updated, and an initial slope a of the characteristic curve is obtained from the matrix Mx. Finally, an offset U0 is calculated from the ratio between the measured voltage U and current I, using the equation U=a*I+U0 The slope is calculated using the initial value a(initial) obtained from the matrix Mx. This offset U0 of the characteristic curve is required for further calculations. The next offset U(n) to be used is set to the calculated value.

[0043] The procedure then continues with a measurement of voltage U and current I. From this, the next offset U(n+1) is calculated. For this, the measured pair of values ​​for voltage U and current I, as well as the slope a of the characteristic curve U = a * I + U(n+1) obtained earlier, can be used, while maintaining the slope a.

[0044] The process then checks whether the newly calculated offset U(n+1) differs from the previously calculated offset U(n) by a value greater than a threshold Dg. If so, an error message is displayed. If not, the slope a is recalculated using the immediately preceding offset U(n) and the voltage U and current I values. The last calculated offset U(n+1) is then used as the new offset U(n). The process then returns to point A, from which voltage U and current I are measured again.

[0045] Furthermore, it is checked whether the stored matrix M1 has already been updated during the current operation of circuit 10. If it has not yet been updated, matrix M1 is updated with the initially measured value pair of voltage U and current I, as well as the first calculated slope a and the first measured temperature T. This makes the stored matrix M1 available as an output value for later commissioning.

[0046] In particular, the procedure described herein allows for advantageous monitoring of a circuit 10 with several power transistors 31, 32, 33, 34, 35, which does not require separate monitoring of the power transistors 31, 32, 33, 34, 35, is therefore easy to implement and has nevertheless proven to be reliable.

[0047] The steps of the method according to the invention can be carried out in the specified order. However, they can also be carried out in a different order, provided this is technically feasible. The method according to the invention can be carried out in one embodiment, for example with a specific combination of steps, in such a way that no further steps are performed. However, further steps can also be carried out in principle, including those not mentioned.

[0048] It should be noted that features may be described in combination in the claims and description, for example to facilitate understanding, even though they can also be used separately. The person skilled in the art recognizes that such features can also be combined independently with other features or combinations of features.

[0049] References in dependent claims may indicate preferred combinations of the respective features, but do not exclude other combinations of features. Reference symbol list 10 Circuit 20 Voltage source 25 additional voltage sources 31, 32, 33, 34, 35 Power transistors 41, 42, 43, 44, 45 resistors 50 electricity meters 51 Shunt resistance 55 voltmeters 60 Temperature sensor 70 Control device Mx Matrix M0 stored matrix M1 stored matrix U voltage I current Temperature a slope U(n) Offset U0 Offset

Claims

[1] Method for monitoring a circuit (10) consisting of several power transistors (31, 32, 33, 34, 35) connected in parallel, using values ​​of one or more parameters of a parameter set of the circuit (10), wherein the method comprises the following steps: - Determine, at successive time points, the respective value of each parameter of the parameter set, and - Determine, for one or more parameters, a deviation between a value of the parameter at a given time and a value of the parameter at a previous time, and - if the deviation is greater than a threshold, an error message will be displayed. [2] Method according to claim 1, - where the parameter set includes as a parameter an offset (U(n)) of a characteristic curve which specifies a voltage (U) across the power transistors (31, 32, 33, 34, 35) as a function of a total current (I) through the power transistors (31, 32, 33, 34, 35). [3] Method according to any one of the preceding claims, - wherein the parameter set includes as a parameter a slope (a) of a characteristic curve which specifies a voltage (U) across the power transistors (31, 32, 33, 34, 35) as a function of a total current (I) through the power transistors (31, 32, 33, 34, 35). [4] Method according to any one of the preceding claims, - wherein an initial value of at least one parameter, or slope (a) of a characteristic curve, is determined based on a matrix (Mx) which contains a voltage (U) across the power transistors (31, 32, 33, 34, 35) and an associated total current (I) through the power transistors (31, 32, 33, 34, 35). [5] Method according to claim 4, - wherein the matrix (M) is obtained when the circuit (10) is put into operation by reading in a stored matrix (M0, M1). [6] Method according to claim 5, - wherein, upon initial commissioning of the circuit (10), a stored matrix (M0) specified during the manufacture of the circuit (10) is read in, and / or - wherein, upon subsequent commissioning of the circuit (10), a matrix (M1) stored during a previous operation of the circuit (10) is read in. [7] Method according to one of claims 5 or 6, - where a matrix to be read (M0, M1) contains temperature-dependent entries, - wherein, prior to the initial determination of the parameter values ​​after commissioning the circuit (10), a temperature (T) is measured and entries based on the measured temperature (T) are used. [8] Method according to any one of claims 5 to 7, - wherein, prior to the initial determination of the parameter values ​​after commissioning the circuit (10), a voltage (U) across the power transistors (31, 32, 33, 34, 35) and an associated total current (I) through the power transistors (31, 32, 33, 34, 35) are measured and written into the matrix (Mx). [9] Method according to any one of claims 4 to 8, - wherein immediately before determining the values ​​of the parameters at a given time, a voltage (U) across the power transistors (31, 32, 33, 34, 35) and an associated total current (I) through the power transistors (31, 32, 33, 34, 35) are measured and written into the matrix (Mx). [10] Method according to any one of claims 4 to 9, - wherein immediately after the first determination of a deviation after commissioning the circuit (10) the matrix (M1) is saved. [11] Method according to any one of claims 4 to 10, - where the matrix (Mx, M0, M1) is stored in a neural network. [12] Method according to any one of claims 4 to 11, - where the circuit (10) is only put into operation if the circuit (10) was out of operation immediately beforehand for at least a specified period of time. [13] Method according to any one of claims 4 to 12, - where only an initial value of a parameter, or the slope (a) of the characteristic curve, is determined based on the matrix (Mx), and - where an initial value of another parameter, or the offset (U(n)) of the characteristic curve, is calculated based on the initial value determined using the matrix (M(x)) and the characteristic curve. [14] Method according to any of the preceding claims, wherein the values ​​of the parameters are determined at least after an initial determination by the following steps: - Measuring a voltage (U) across the power transistors (31, 32, 33, 34, 35) and an associated total current (I) through the power transistors (31, 32, 33, 34, 35), - Calculating a new value of a first parameter, or an offset (U(n)) of the characteristic curve, based on the measured voltage (U) and the measured associated total current (I), and - Calculating a value of a second parameter, or a slope (a) of the characteristic curve, based on the measured voltage (U) and the measured associated total current (I) as well as the previous value of the first parameter. [15] Method according to any one of the preceding claims, - wherein the steps to be performed during operation of the circuit (10) are repeated several times. [16] Method according to any one of the preceding claims, - where the deviation for one, several or all parameters is determined as an absolute deviation. [17] Method according to any one of the preceding claims, - where the deviation for one, several or all parameters is determined as a relative deviation. [18] Method according to any one of the preceding claims, - where the power transistors (31, 32, 33, 34, 35) are MOSFETs. [19] Control device (70) configured to perform a method according to any of the preceding claims. [20] Circuit (10), comprising - several power transistors (31, 32, 33, 34, 35), and - a control device (70) according to claim 19.

Images