Method and system for sensor monitoring in a modular electrical system

Abstract

The invention relates to a method and a system for sensor monitoring in a modular electrical system (1), which comprises several modules (M1, M2, ..., MN) each with a source sub-module (2) and a switching sub-module (3) for activating or deactivating the module (M1, M2, ..., MN), comprising the steps: a) Setting at least one system configuration state, b) Acquiring a state-specific system voltage (V_o) with a system-specific voltage sensor (4) for the at least one system configuration state, c) Determining at least one sensor parameter of the system-specific voltage sensor (4) and / or detecting a malfunction of the system-specific voltage sensor (4) as a function of the state-specific system voltage (V_o).

Description

[0001] The invention relates to a method and a system for sensor monitoring in a modular electrical system.

[0002] Modular electronics has immense potential, but its complexity also complicates monitoring and protection functions. Extensive research exists on the development and design of monitoring circuits and functions for electronic circuits, including in the automotive industry. However, measurements / sensors remain the weak point in these systems, as they require lengthy commissioning procedures and costly regular maintenance.

[0003] Due to manufacturing tolerances, the sensors, as a key component of a monitoring system, exhibit inherent differences. Furthermore, the components of the monitoring system are also susceptible to distortion, parameter drift, aging, and failure, which can significantly impair the overall performance of the system.

[0004] The scalability and enormous potential of modular electrical systems come at the cost of increased complexity and stricter requirements for stable and safe operation. As modular, reconfigurable structures such as storage / battery systems (and photovoltaic systems) become increasingly prevalent, the need for accurate and reliable sensor measurements is crucial to ensuring their optimal performance and safety.

[0005] However, there are several challenges associated with addressing sensor-related problems in these systems: • Unsafe sensor behavior: Sensor inaccuracies can occur due to manufacturing variations, aging, environmental factors, or damage, leading to unsafe behavior and potentially affecting the system's performance and safety. • Balancing module interchangeability and maintenance: The possibility of exchanging, removing, or adding modules to these systems to adapt to new requirements makes the maintenance of these systems problematic, as each added or replaced module requires a new commissioning and recalibration process. • Compromise between sensor cost and accuracy: High-precision sensors are generally more expensive, while less expensive sensors often provide less accurate measurements. Finding a balance that minimizes costs while ensuring adequate system performance and safety is a challenge. • Gradual divergence: Over time, sensor inaccuracies can lead to a gradual deviation of measured values ​​from actual values, affecting overall system performance and potentially causing undetected problems. Furthermore, the dynamics of the sensors and their parameters can deviate from their nominal values ​​over time or with changes in operating conditions. • Cross-sensitivity and influence of environmental conditions: Sensors are generally dependent not only on the measured quantity but also on various other influences. For example, almost every sensor is also temperature-dependent. Correcting such cross-sensitivities usually requires intensive characterization of the sensors for a multitude of influencing factors in order to describe their behavior mathematically, and potentially additional, and possibly costly, sensors in a product to make corrections based on the other sensor signals. Furthermore, some cross-sensitivities are unknown and can therefore uncontrollably alter the sensor's output signal in an undesirable way. • Malfunctions and failures: Undetected sensor faults can lead to system failures or safety risks, resulting in unplanned downtime, increased maintenance costs, and potential damage to equipment or the environment.

[0006] US9,325,282 B2 applies to integrated circuits in general and, in particular, to integrated circuits that use circuit topologies that enable higher manufacturing yields.

[0007] US11,320,810 B2 concerns the field of autonomous or semi-autonomous devices, in particular multimodal sensor technology in autonomously driving vehicles with self-healing capabilities.

[0008] Other known solutions for sensor monitoring consist of redundant measurements. This is described, for example, in US2020 / 397546 A1. However, this is very expensive and requires high bandwidth for signal transmission. A disadvantage is that distortions cannot be corrected with this method; it is sensitive to gradual deviations in the sensor or model parameters and depends on the accuracy of the model.

[0009] The document “S. Dey, S. Mohon, P. Pisu and B. Ayalew, “Sensor Fault Detection, Isolation, and Estimation in Lithium-Ion Batteries,” in IEEE Transactions on Control Systems Technology, vol. 24, no. 6, pp. 2141-2149, Nov. 2016, doi: 10.1109 / TCST.2016.2538200” describes a model-based approach to estimate voltages on modules.

[0010] The document “J. Lu, Y. Hu, G. Chen, Z. Wang and J. Liu, “Mutual Calibration of Multiple Current Sensors With Accuracy Uncertainties in IPMSM Drives for Electric Vehicles,” in IEEE Transactions on Industrial Electronics, vol. 67, no. 1, pp. 69-79, Jan. 2020, doi: 10.1109 / TIE.2019.2896320” describes an approach to the regular maintenance of modular electrical systems. Frequent calibration and maintenance can detect and correct gradual sensor distortions and errors. This method has the disadvantage that it can potentially hinder system operation, especially if the system needs to be shut down; it requires trained personnel for high-voltage systems; it incurs high costs; and it must be repeated after every change, such as when replacing a module.

[0011] The document “K. Zhang, X. Hu, Y. Liu, X. Lin and W. Liu, “Multi-fault Detection and Isolation for Lithium-Ion Battery Systems,” in IEEE Transactions on Power Electronics, vol. 37, no. 1, pp. 971-989, Jan. 2022, doi: 10.1109 / TPEL.2021.3098445” describes nested measurements between battery cells in a series circuit. However, it does not allow for distortion correction, is sensitive to gradual changes in parameters or behavior, and is difficult to apply when replacing or adding modules.

[0012] The document “X. Hu, K. Zhang, K. Liu, X. Lin, S. Dey and S. Onori, “Advanced Fault Diagnosis for Lithium-Ion Battery Systems: A Review of Fault Mechanisms, Fault Features, and Diagnosis Procedures,” in IEEE Industrial Electronics Magazine, vol. 14, no. 3, pp. 65-91, Sept. 2020, doi: 10.1109 / MIE.2020.2964814” describes data-driven methods for fault diagnosis. However, it requires extensive measurement datasets that are generally difficult to obtain, exhibits low reliability and unclear behavior, is sensitive to gradual parameter drift in the modules and sensors as well as to distortions, and has poor generalization properties; in particular, it must be retrained after every change.

[0013] The document N. Tashakor, J. Kacetl, D. Keshavarzi and S. Goetz, “Topology, Analysis, and Modulation Strategy of a Fully Controlled Modular Reconfigurable DC Battery Pack With Interconnected Output Ports for Electric Vehicles,” in IEEE Transactions on Transportation Electrification, vol. 10, no. 1, pp. 1180-1193, March 2024, doi: 10.1109 / TTE.2023.3263775 describes modular battery-integrated converters or so-called dynamically reconfigurable battery packs.

[0014] The technical problem therefore arises of enabling reliable determination of sensor parameters in modularly designed electrical systems, in particular to compensate for sensor distortions and / or to enable reliable fault detection.

[0015] The solution to the technical problem is provided by the articles with the features of the independent claims. Further advantageous embodiments of the invention are described in the dependent claims.

[0016] A method for sensor monitoring, in particular voltage sensor monitoring, in a modular electrical system is proposed. The method can be used for sensor calibration and / or fault detection.

[0017] A modular electrical system can refer to an electrical system consisting of several largely uniform modules. Modules in such a system can typically be added or removed with minimal impact on the overall design. Modular reconfigurable systems generally consist of modules with actively controllable elements (electrical, including electronic, or mechanical switches or relays) that allow the overall architecture (e.g., how the modules form the larger subsystem) to be modified and the system adapted or extended with (new) functions. Modular electrical systems exhibit high flexibility, scalability, and adaptability. In particular, a modular electrical system can be a power electronic system, especially for energy storage and (renewable) energy generation. Reconfigurable modular electrical storage systems (e.g.,Batteries advantageously allow for dynamic adaptation of the storage element configuration to varying energy requirements and operating conditions. These systems can consist of individual modules that can be connected in series or parallel, enabling voltage and current to be adjusted as needed. This adaptability increases the overall performance, efficiency, and lifespan of the energy storage system while simultaneously reducing costs, as it facilitates the replacement of individual modules rather than the entire system (e.g., the battery pack).

[0018] Modular reconfigurable photovoltaic (PV) systems and supercapacitors are another application of modular electrical systems. These systems can consist of individual solar modules that can be connected in various configurations to maximize energy yield under different environmental conditions, such as variable sunlight intensity or partial shading. By allowing real-time adjustments to the system configuration, modular reconfigurable PV systems can improve the overall efficiency, resilience, and reliability of solar power generation.

[0019] Another example of a modular electrical system is a system for current or voltage conversion, in particular multilevel converters.

[0020] Generally speaking, modular and reconfigurable electrical systems consist of several modules, particularly interchangeable ones, which can be easily reconfigured to adapt to changing operating conditions, requirements, or environments. In their most basic structure, the modules can be switched or bridged. The modular electrical system comprises several modules, each with a source sub-module. The source sub-module serves to provide electrical energy and can, in particular, include at least or exactly one voltage or current source. A voltage source can be designed as a battery cell or accumulator. However, it is also conceivable that the voltage source could be designed as a photovoltaic cell or in another form. A module further includes a switching sub-module for activating or deactivating the module.

[0021] In other words, a module can comprise at least one electrical source or energy storage element, with the possibility of cascading two or more of these elements. Furthermore, it can include a configurable switching device that allows the module to dynamically change its electrical connection within the system, in particular to form a series or parallel connection or a quasi-DC-DC stage (where the module can contribute a fraction of its voltage or power to the overall system) with at least one other module. The switching device also allows for a bypass configuration in which the corresponding module is not an active part of the system.

[0022] The switching device can consist of electronic components (e.g., MOSFETs, IGBTs, FETs) or mechanical elements (e.g., relays) arranged in topologies such as single switches, half-bridges, full-bridges, and / or combinations thereof. Such components / elements are described, among other places, in the document by N. Tashakor, J. Kacetl, D. Keshavarzi, and S. Goetz, "Topology, Analysis, and Modulation Strategy of a Fully Controlled Modular Reconfigurable DC Battery Pack With Interconnected Output Ports for Electric Vehicles," in IEEE Transactions on Transportation Electrification, vol. 10, no. 1, pp. 1180–1193, March 2024, doi: 10.1109 / TTE.2023.3263775. Reference can therefore be made to the disclosure content of the corresponding document.

[0023] Inductive and / or capacitive components can be connected between the modules and between the source and switching submodules or components of these submodules, e.g. to support transient response or filtering.

[0024] Each module can have selected measurement points on the source submodules or components of such a submodule and / or on the switching submodules or components of such a submodule, in particular an output terminal. However, it is conceivable that some modules may not have any measurement points and that module-specific parameters may be determined based on measurements at other modules or at the system level.

[0025] By setting an activated switching state, the module, and in particular the source submodule, can be activated. In the activated state, it is part of the modular electrical system. By setting a deactivated switching state, the module, and in particular the source submodule, can be deactivated. In the deactivated state, the module is not part of the electrical system. In other words, the switching submodule can be used to change how the source submodule is integrated into the electrical system.

[0026] If the modular electrical system comprises, for example, a series circuit, the module can be connected in series with other modules in the series circuit when activated. When deactivated, it can be bypassed. If the modular electrical system comprises, for example, a parallel circuit, the module can be connected in parallel with other modules in the parallel circuit when activated. When deactivated, it can be disconnected from the parallel circuit. The modular electrical system can comprise at least one series circuit and / or at least one parallel circuit of modules.

[0027] Each module therefore comprises at least one type of energy storage / source or conversion unit (ES or EC) and an electronic (or even mechanical) component that can change the state through which the ESS is integrated into the overall structure. The switching elements of the switching submodule can be mechanical or electronic components, ideally using MOSFETs, IGBTs, thyristors, GANs, and SiCs.

[0028] The modular electrical system comprises at least one system-specific voltage sensor. This sensor is used in particular to detect the system's output voltage. Preferably, but not necessarily for at least one embodiment, the system comprises at least one module-specific voltage sensor for each module. This sensor is used in particular to detect the output voltage of the corresponding module.

[0029] It is also possible that the modular electrical system includes a system-specific current sensor, which serves in particular to detect an input or output current of the system, and / or includes at least one module-specific current sensor for each module, which serves in particular to detect an output voltage of the corresponding module.

[0030] It is also possible that the modular electrical system includes at least one temperature sensor for each module, which serves in particular to detect the temperature of the corresponding module.

[0031] A sensor can be connected to a control and evaluation unit of the system via data and / or signal transmission. It is conceivable that the sensors are connected to a module-specific control and evaluation unit, which in turn is connected to the system-specific control and evaluation unit described above. A control and evaluation unit can be designed as a computing unit or include one, and a computing unit can be designed as a microcontroller or integrated circuit, or include one or more of these.

[0032] The modular electrical system can, for example, comprise a number N of (de)activatable modules. These modules can be used to set a number M of different system output voltage levels. Preferably, the output voltage level of each module is the same, or the levels do not deviate from each other by more than a predetermined amount, e.g., 10% or 5% of an average module output voltage level. In this case, the number M of levels can correspond to the number N of modules, particularly if all modules can be connected in series.

[0033] Preferably, no filter element for voltage filtering is arranged between the modules and the system output terminals.

[0034] The process includes the following steps: a) Setting at least one system configuration state, b) Capturing a state-specific system voltage with a system-specific voltage sensor for at least one system configuration state, c) Determining at least one sensor parameter of the system-specific voltage sensor and / or detecting a sensor malfunction as a function of the state-specific system voltage. Sensor monitoring is performed by determining the at least one sensor parameter and / or detecting the sensor malfunction.

[0035] Setting at least one system configuration state involves setting the enabled or disabled state of modules. This allows an electrical system with predetermined properties to be provided. Exemplary system states are explained in more detail below. In its simplest form, all modules can be disabled, thus creating an open-circuit state for the system. At this operating point, the system's open-circuit voltage is then measured as the state-specific system voltage.

[0036] Then, for example, a distortion (bias) of the system-specific voltage sensor at the zero voltage point can be determined as this open-circuit voltage as a sensor parameter. Other sensor parameters that can be determined using the proposed method include aging or a change in the distortion or aging. After determining the sensor parameter, an output signal of the sensor can be determined as a function of the sensor parameter. Thus, a method for determining a sensor output signal is also described, which includes the explained determination of at least one sensor parameter as well as the determination of the sensor output signal as a function of the sensor parameter thus determined.

[0037] Alternatively or cumulatively, a sensor malfunction of the system-specific voltage sensor can be detected if the open-circuit voltage deviates from a predetermined voltage level, e.g., zero voltage, by more than a predetermined amount. Preferably, however, and as explained in more detail below, a variety of different system configuration states are set, and the determination of the sensor parameter and / or the detection of the sensor malfunction is performed depending on the state-specific system voltage. If a malfunction is detected, an error signal can be generated. This can, for example, be transmitted to a higher-level system and / or output to a user in a perceptible manner, e.g., visually or audibly.

[0038] Preferably, a module-specific voltage sensor is also used to detect the module voltage of at least one module, preferably of each module. Then, the at least one sensor parameter of the system-specific voltage sensor can be determined as a function of the at least one module voltage.

[0039] This advantageously enables the reliable determination of sensor parameters in modular electrical systems, particularly for calibration or error compensation, and / or reliable fault detection, especially for more complex and versatile systems where there is an increasing need for accurate and reliable measurements to ensure optimal performance and safety. The method is particularly applicable to modular and therefore reconfigurable storage / battery systems, modular reconfigurable PV systems, and modular energy conversion systems.

[0040] This approach utilizes the fundamental idea of ​​setting specific system configurations, which offers a unique advantage in detecting sensor errors and / or correcting distortions in particular. This can be achieved primarily by independently monitoring module and system output measurements and by leveraging the available degrees of freedom for switching operations to reconfigure the module arrangement within the system. As explained in more detail below, this method can be implemented in various system operating modes: • Disconnect or Idle (idle state): The system is disconnected from any load and / or chargers, therefore the system can be easily reconfigured to bring the operating mode into the desired range. • Under load / charge (load state): The system actively supplies or draws power. Because the system's output voltage, current, or power is controlled, the system can only be reconfigured as long as the same output levels are kept relatively constant. Therefore, the number of available redundant states is rather limited.

[0041] In addition to determining a sensor parameter of the system-specific voltage sensor and / or diagnosing its fault, the procedure can also be applied for the same purpose to the aforementioned module-specific voltage sensors.

[0042] In a further embodiment, in a system configuration state, a module voltage from at least one module is additionally detected using a module-specific voltage sensor, wherein the at least one system-sensor-specific sensor parameter is additionally determined as a function of the at least one module voltage. For example, it is possible that exactly one module is activated. The detected state-specific system voltage can then be compared with the detected module voltage. For example, a distortion as the difference between the two voltages can then be determined as a system-sensor-specific parameter, particularly if it is assumed that the module-specific voltage sensor has a higher measurement accuracy, e.g., a lower tolerance. In particular, the distortion can thus be determined for at least one operating point in which the module voltage of a module is provided as the output voltage.

[0043] Alternatively or cumulatively, a malfunction of the system-specific voltage sensor can be detected if the detected voltages deviate from each other by more than a predetermined amount.

[0044] This advantageously further improves the reliability and accuracy in determining sensor parameters and / or the reliability of fault detection.

[0045] In a preferred embodiment, a sensor parameter of the module-specific voltage sensor is determined as a function of the state-specific system voltage and the module voltage. For example, it is possible that exactly one module is activated. The detected state-specific system voltage can then be compared with the detected module voltage. A distortion, defined as the difference between the two voltages, can then be determined as a module-specific parameter, particularly if the system-specific voltage sensor has a higher measurement accuracy. Alternatively or cumulatively, a malfunction of the system-specific voltage sensor can be detected if the detected voltages deviate from each other by more than a predetermined amount.

[0046] This advantageously further improves the reliability and accuracy in determining sensor parameters and / or the reliability of fault detection, as module-specific voltage sensors can also be monitored.

[0047] In a further preferred embodiment, a plurality of different system configuration states are set and the corresponding state-specific system voltages are recorded, wherein the at least one system sensor-specific sensor parameter is determined as a function of the recorded system voltages. Naturally, the module-specific module voltages of the activated modules can also be recorded in these different system configuration states and taken into account when determining the system sensor-specific sensor parameter. Alternatively or cumulatively, the recorded voltages can also be taken into account when determining the module sensor-specific sensor parameters. The same applies to fault detection.

[0048] In a modular electrical system where N modules can be connected in series, N system configuration states can be set, in each of which exactly one module is activated and the remaining modules are deactivated. The number of system configuration states corresponds to the number of modules. In each of these N system configuration states, a different module is activated. If the modules have the same module output voltage level, different system configuration states can be set to provide an initial system output voltage level. In each system configuration state, a difference between the state-specific system voltage and the respective module voltage can then be determined. An average of these differences can then be calculated. Finally, a distortion can be determined as a system sensor-specific parameter based on this average.It is possible that module voltages deviating from the state-specific system voltage recorded in the corresponding system configuration state by more than a predetermined amount (e.g., more than 10% or more than 5%) will not be considered in this determination. In such cases, the averaging must be adjusted accordingly. Thus, for an operating point where the system provides a system output voltage with a first system output voltage level, a distortion of the system-specific voltage sensor can be determined.

[0049] Depending on the number of adjustable system output voltage levels, this procedure can be repeated for additional or even all adjustable system output voltage levels to determine any distortion of the system-specific voltage sensor for further operating points of the system. For example, to provide a second system output voltage level, different module pairs can be activated, each consisting of two modules connected in series.

[0050] If the distortion is determined for multiple operating points, a distortion correction function can also be determined, which enables the determination and thus also the correction of distortions, especially for unknown operating points. This function can be determined, for example, using linear or nonlinear regression.

[0051] This advantageously further improves the reliability and accuracy in determining sensor parameters and / or the reliability of fault detection.

[0052] In another embodiment, the system configuration state is defined as follows: a) a first system configuration state is set in which all modules are disabled or b) a second system configuration state is set in which exactly one module of the system is activated or c) a third system configuration state is set in which exactly a number n of the system modules are enabled, where the number n is greater than one.

[0053] These system configuration states allow for the advantageous provision of different system output voltage levels, thus enabling the determination of distortion for different operating points.

[0054] This also advantageously improves the reliability and accuracy in determining sensor parameters and / or the reliability of fault detection.

[0055] In a further embodiment, a state-specific system voltage and a module voltage of the respective activated module are determined for at least one second system configuration state, wherein the at least one sensor parameter of the system-specific voltage sensor is determined as an updated sensor parameter depending on this state-specific system voltage and the module voltage. It is assumed here that the state-specific system voltage used to determine the updated sensor parameter was previously determined using an unupdated and potentially incorrect sensor parameter.

[0056] For example, the module whose voltage sensor has the highest measurement accuracy can be activated. In this particular case, the modular electrical system can be designed such that the voltage sensor of exactly one or more modules has a higher measurement accuracy than the measurement accuracy of the voltage sensors of the remaining modules.

[0057] Of course, the state-specific system voltage and a module voltage of the respective activated module can be determined for several system configuration states and / or the state-specific system voltage can be determined for the first system configuration state, whereby the at least one sensor parameter of the system-specific voltage sensor is determined as an updated sensor parameter and / or the system voltage determined in the first system configuration state, depending on these state-specific system voltages and the respective module voltage.

[0058] Then, for several different second system configuration states, the module voltage of the respective activated module and the state-specific system voltage can be determined, whereby the state-specific system voltage is determined with the updated sensor parameter, wherein the at least one sensor parameter of the module-specific voltage sensor is determined as a function of the respective module voltage and the state-specific system voltage.

[0059] In other words, with comparatively little effort, the sensor parameter of the system-specific voltage sensor is first determined, and this is then used to determine the sensor parameters of the module-specific voltage sensors.

[0060] This advantageously achieves a rapid and easily implementable improvement in the reliability and accuracy of determining sensor parameters and / or the reliability of fault detection.

[0061] In a further embodiment, at least one sensor parameter is determined as a function of at least one sensor tolerance. In particular, the lower the sensor tolerance or the higher the respective measurement accuracy, the more the detected (system or module) output voltages can be taken into account when determining sensor parameters. This also advantageously improves the reliability and accuracy in determining sensor parameters and / or the reliability of fault detection.

[0062] In a further embodiment, the method is carried out in an idle state or under load. These states have already been explained. However, under load, only those system configuration states can be set that provide a load-dependent system output voltage or in which the provided system output voltage does not deviate from this load-dependent system output voltage by more than a predetermined amount. While in the idle state a sensor parameter can advantageously be determined for a multitude of operating points, or a malfunction of the system-specific voltage sensor can be detected for a multitude of operating points, carrying out the method under load advantageously increases the availability of the sensor monitoring.

[0063] In another embodiment, the modules are activated with a time delay to set different system configuration states. For example, to provide a system output voltage corresponding to the sum of the output voltages of two activatable modules, a first module and then a second module can be activated. This advantageously reduces the magnitude of voltage fluctuations when setting the different system configuration states, which on the one hand reduces the system load and on the other hand leads to faster transient responses. This, in turn, can increase the measurement accuracy of the voltage sensing.

[0064] The time-delayed activation of a module can be achieved, in particular, by using carrier signals for pulse-width modulated activation of different modules with a predetermined phase shift. In such pulse-width modulated activation, the carrier signal, which can be, for example, a triangle or sawtooth wave, is compared to a constant reference signal. If the carrier signal rises above the reference signal, a module is activated. If the carrier signal falls below the reference signal, the module is deactivated. For pulse-width modulated activation, the reference signal can be set to a voltage-level-dependent (constant) value to adjust different system output voltage levels. This advantageously results in a simple implementation of the described method.

[0065] Furthermore, a voltage measurement can be performed for a predetermined time period after a switching signal edge to activate or deactivate a module, or for a predetermined time period after reaching a predetermined amplitude of the carrier signal.

[0066] A further proposal is a system for sensor monitoring in a modular electrical system, comprising the modular electrical system, a device for setting at least one system configuration state, a system-specific voltage sensor, and at least one control and evaluation device, wherein the system is configured to perform a method according to one of the embodiments described in this disclosure. Naturally, the system can also include the described module-specific voltage sensors. Thus, the system offers the technical advantages already explained. The control and evaluation device can constitute the device for setting at least one system configuration state.

[0067] A vehicle with such a system is further described, whereby the modular electrical system in this case can be, in particular, a storage system or a converter system.

[0068] The invention is explained using exemplary embodiments. The figures show: Fig. 1. A schematic view of a modular electrical system, Fig. 2 a schematic view of a modular electrical system in a further embodiment, Fig. 3 a schematic view of a modular electrical system in a further embodiment, Fig. 4 a schematic flowchart of a method according to the invention, Fig. 5 a schematic flowchart of a method according to the invention in a further embodiment and Fig. 6 probability distribution functions of a system-specific voltage sensor and a set of module-specific voltage sensors.

[0069] In the following, identical reference symbols denote elements with the same or similar technical characteristics.

[0070] Fig. Figure 1 shows a schematic view of a modular electrical system 1. The system 1 comprises a number N of modules M1, M2, ..., MN. Each module M1, M2, ..., MN includes an energy storage or energy source sub-module 2 and a switching sub-module 3 for activating and deactivating the respective module M1, M2, ..., MN. In the Fig. In the illustrated embodiment 1, the modules M1, M2, ..., MN can be connected in series. An activated module M1, M2, ..., MN is part of the series connection, while a deactivated module is bypassed. Various system configuration states can be set by controlling the switching submodules 3, whereby the number and / or the sequence number(s) of the activated modules M1, M2, ..., MN vary in different system configuration states. The modules M1, M2, ..., MN can be interchangeable. The illustrated system 1 allows for easy reconfiguration, particularly to adapt to changing operating conditions, requirements, or environments. The output terminals T1, T2 of system 1 are also shown. A state-specific system voltage can be detected at these terminals.

[0071] Fig. Figure 2 shows a schematic view of a modular electrical system 1 in a further embodiment. The system 1 comprises N modules M1, M2, ..., MN. The system 1 further comprises a system-specific voltage sensor 4, which detects a system voltage drop between output terminals T1, T2 of the system 1. The system 1 also includes, for each module M1, M2, ..., MN, a module-specific voltage sensor 5_1, 5_2, ..., 5_N, each of which detects an output voltage of one of the modules M1, M2, ..., MN. The system 1 can optionally also include a system-specific current sensor 6 for detecting the current flowing into or out of the system 1 via the first output terminal. Furthermore, the system 1 can optionally also include current sensors 7_1, 7_2, ..., 7_N, which detect a module current.

[0072] The sensors 4, 5_1, 5_2, ..., 5_N, 6, 7_1, 7_2, ..., 7_N are connected via a signal or data line 8 to a control and evaluation unit 9. This unit serves as a sensor monitoring device and can determine a sensor parameter, in particular a distortion, of a voltage sensor 4, 5_1, 5_2, ..., 5_N and / or detect a malfunction of a voltage sensor 4, 5_1, 5_2, ..., 5_N.

[0073] Fig. Figure 3 shows a schematic view of a modular electrical system 1 in a further embodiment. The in Fig. The system shown in section 3 is essentially the same as the one in Fig. System 1 as depicted in Figure 2 is formed. Therefore, reference can be made to the corresponding explanations. In contrast to the system shown in Figure 2, the system is formed in Figure 2. Fig. In the embodiment shown in Figure 2, sensor output signals from the module-specific sensors 5_1, 5_2, ..., 5_N, 7_1, 7_2, ..., 7_N are transmitted via signal or data lines to module-specific evaluation units 10_1, 10_2, ..., 10_N and from these also via a signal or data line to the control and evaluation unit 9.

[0074] It is assumed that in the Fig. 1, Fig. 2 and Fig. If the module output voltage levels of each module M1, M2,..., MN in the three depicted systems 1 are equal, then a number N of different system output voltage levels can be set. Furthermore, a method for sensor monitoring of the modular electrical system 1 can be implemented. For this purpose, the control and evaluation unit 9 sets several different system configuration states. In each of these states, a state-specific system voltage is measured by the system-specific voltage sensor 4, and the module voltages of the modules activated in the set system configuration state are measured by the module-specific voltage sensors. Then, a sensor parameter of the system-specific voltage sensor 4 and the module-specific voltage sensors 5_1, 5_2,..., 5_N, in particular a bias, is determined by the control and evaluation unit 9. Alternatively, a malfunction of the voltage sensors 4, 5_1, 5_2,..., 5_N can be detected.

[0075] To determine a distortion of the system-specific voltage sensor 4, the following steps can be performed in an idle state: i. In a first step, to establish an initial system configuration state, all modules M1, M2,..., MN are bypassed. The module voltages are then zero, and the system output voltage V_oO is measured. Since all modules are bypassed, the target system output voltage is zero. The distortion of the system-specific voltage sensor 4 at the operating point of system 1, where a zero voltage is provided, is B_0 = V_oO. ii. Then, various system configuration states are set to provide an initial system output voltage level, which corresponds to the module voltage of exactly one activated module M1, M2, ..., MN. In these states, exactly one module M1, M2, ..., MN is activated, while the remaining modules M1, M2, ..., MN are deactivated, i.e., bypassed. The system output voltage V_o1,i and the module voltage of the activated i-th module V_i are then measured. Module voltages V_i that deviate from the corresponding system output voltage Vo1,i by more than a predetermined amount are disregarded. The difference between the considered voltages is then determined, and an average of these differences is calculated. This average is the distortion B_1,avg of the system-specific voltage sensor 4 at the operating point of system 1, in which the initial system output voltage level is provided. B1,avg=∑i(Vo1,i−Vi)(N−noutliers) where n_outlier denotes the number of module voltages not considered. iii. This process is repeated similarly, this time setting different system configuration states to provide a second system output voltage level, which corresponds to the sum of the module voltages of two activated modules M1, M2, ..., MN. Thus, in the respective system configuration states, a first module i and a second module j are activated. Again, the difference between the respective system output voltage V_o2,ij and the respective module voltages V_i, V_j, and their average value, are determined. This average value is the distortion B_2,avg of the system-specific voltage sensor 4 at the operating point of system 1, in which the second system output voltage level is provided. B2,avg=2∑j∑i(Vo2,ji−Vi−Vj)(N−noutliers)(N−1−noutliers)with i≠j=1,…,N and i <J where n_outlier denotes the number of module voltages not considered. iv. This can then be repeated analogously for all adjustable system output voltage levels to determine distortions for further operating points.

[0076] A continuous distortion correction function for continuous voltage measurement can then be derived by linear and nonlinear multivariable regression, for example, by fitting a polynomial function P(x) = a_0 + a_1·x + a_2·x^2 + ... + a_q·x^q of degree q to the average distortions at each operating point, where x is the number of activated modules M1, M2,..., MN or the system output voltage, and P(x) is the distortion. The coefficients a_0, a_1,..., a_q can be determined by applying a regression method such as the method of least squares. Once the polynomial function is fitted, it can be used to correct the system-level measurements for each voltage level within the specified range, with V_o_corrected = V_o - P(x).

[0077] Therefore, the system output voltage can be corrected using the distortions determined as explained.

[0078] When setting a system configuration state, sufficient time should elapse between the switching operation and the measurement to allow the sensor measurements to reach a stable state, which depends on the dynamics of sensor 4, 5_1, 5_2, ..., 5_N and system 1, e.g. at least a few µs, preferably a few ms, even better a few seconds.

[0079] To reduce the noise for each system output voltage level, the higher-frequency components of the measurements, preferably harmonics that are more than one hundred times the switching frequency, preferably more than twenty times, and ideally more than five times, should be filtered out by a filter device. The appropriate filter device can be part of System 1.

[0080] To determine the distortion of a module-specific voltage sensor 5_1, 5_2, ..., 5_N in the open-circuit state, in each system configuration state i, in which, for example, a first system output voltage level is provided, one of the modules M1, M2, ..., MN, e.g., the first module M1, can be activated while the other modules M2, ..., MN are bypassed. The difference between the measurements V_o1,i, V_k of the system-specific voltage sensor 4 and the corresponding module-specific voltage sensor 5_1 yields the distortion of the module-specific voltage sensor 5_1 according to B_5_1,1 = V_o1,i - V_k - B1,avg, where B1,avg is the pre-determined distortion of the system-specific voltage sensor 4. Sufficient time should elapse between setting the individual system configuration states to ensure that the measured voltages are sufficiently noise-free.

[0081] Under load conditions, modules M1, M2, ..., MN can be switched into or out of system 1 as desired, as long as the performance of system 1 remains relatively unaffected. Preferably, there is no filter between the system-specific voltage sensor 4 and the voltage generated by system 1, and the output voltages of all modules M1, M2, ..., MN have relatively similar voltage values ​​and differ from each other by less than 10%, ideally less than 5%.

[0082] The following steps can then be carried out to determine and correct the distortion of the system-specific voltage sensor 4. i. In a first step, all module combinations that result in similar system output voltage levels are determined. For example, if the number of modules connected in series is x and the total number of modules is N, there can be as many permutations as n=N! / ((N - x)!), provided the output voltages of modules M1, M2, ..., MN are approximately equal. ii. Then all system configuration states i defined in this way are traversed and the sum of the module voltages V_M,i of the activated modules is determined in each case and compared with the state-specific system voltage V_o,i, whereby the distortion Bias_i of the system-specific voltage sensor 4 for the output voltage provided in the system configuration state is determined as the difference between the state-specific system voltage V_o,i and the sum of the module voltages V_M,i. iii. The distortions thus determined are then averaged to determine the distortions of the system-specific voltage sensor 4 for this specific system output voltage level. The averaged distortion is given by Bias_avg=(1 / (n−noutliers))∑((i=1 to(n−noutliers)))Bias_i where n_outliers is the number of cases in which the calculated biases deviate significantly, e.g. by more than 10%, from a mean of all biases Bias_i. iv. The calculated distortion correction point is then used - as explained above - to update the distortion correction function.

[0083] At high loads, where the voltage drops on the conductor paths and the switching submodules 3 may be higher than a predetermined threshold, these can also be included in the calculation of the distortions, as described, for example, in the document “Tashakor et al., “Low-Cost Parameter Estimation Approach for Modular Converters and Reconfigurable Battery Systems Using Dual Kalman Filter”, in IEEE Transactions on Power Electronics, vol. 37, no. 6, pp. 6323-6334, June 2022, doi: 10.1109 / TPEL.2021.3137879”.

[0084] In cases where the modules M1, M2, ..., MN have significantly different module voltages (e.g., in applications known as asymmetric modular multilevel converters), the described method is also applicable. However, the number of connected modules M1, M2, ..., MN depends on the individual module voltages and must be calculated on a case-by-case basis.

[0085] To determine and correct the distortion of the module-specific voltage sensors 5_1, 5_2, ..., 5_N, two variants can be implemented under load: In a first variant, if only one module M1, M2, ..., MN is activated at a time, a voltage jump detected with the distortion-corrected system-specific voltage sensor 4 can be interpreted as the result of switching on the module M1, M2, ..., MN in system 1 and the amplitude of the voltage jump is a good approximation of the module voltage of the module in question.

[0086] If the module j has changed its state at a time t and the resulting voltage jump in the measurements with the system-specific voltage sensor 4 results in ΔV_o=V_o(t)−V_o(t−1) If this results in the following distortion of the measurement at module j, then the distortion is as follows: B_Mj(t)=V_j(t−1)−ΔV_o

[0087] Normally, measurements under load contain a high level of noise, which is why a more practical approach is to use a stepwise update procedure as follows: B_(Mj)(t)=f(B_(Mj)(t))+(V_j(t−1)−ΔV_o) where f(x) + g(x) = x, and f(x) and g(x) have slow dynamics to reduce the effects of measurement noise. For example: f(x) = (1 - k)x and g(x) = kx, where k << 1. In principle, the combination of f(x) and g(x) provides a low-pass filter. Assuming that S is a vector of module states and B_ML is a vector of distortions, this can be expressed more generally as follows: BML(t)=(I−S(t)+S(t−1)).BML(t)+(S(t)−S(t−1)).(f(BML(t))+g(VML(t−1)−ΔVo)) where V_ML (t) is the vector of module-specific voltage measurements at time t.

[0088] Referring to this nomenclature, in a second variant all elements of the bias vector can be updated simultaneously by B_ML(t)=(I−S(t))B_ML(t)+S(t)(f(B_ML(t))+g(S(t)T V_ML(t)−V_o))

[0089] The functions f(x) and g(x) behave as noise reduction and limiting functions and have the following properties. f(x)+g(x)=x,g(x)={0|x|>>0h(x)else,h(x)≤x.

[0090] In a further implementation of distortion correction under load, to avoid fixed estimates of distortion due to a fixed sequence of permutations when connecting modules M1, M2, ..., MN, the phase shift of the carrier signals and / or the connection sequence of modules M1, M2, ..., MN can also be changed after a certain period of time, as explained below.

[0091] A malfunction of the system-specific voltage sensor 4 can be detected by observing fluctuations in the measured values ​​when the module configurations or system configuration states change, or when the measured value acquired by the system-specific voltage sensor 4 varies even though no change is made to the system configuration state and the module voltages acquired by the module-specific voltage sensors 5_1, 5_2, ..., 5_N show no changes. Thus, a malfunction can be detected by observing the measured values ​​acquired by the system-specific voltage sensor 4 when a system configuration state is changed. If a trend of the measured values ​​shows unexpected fluctuations or inconsistencies, a malfunction can be detected.

[0092] A malfunction of a module-specific voltage sensor 5_1, 5_2, ..., 5_N can be detected if a system configuration state is changed and the measured value profile recorded by the corresponding sensor 5_1, 5_2, ..., 5_N shows unexpected fluctuations or inconsistencies. A malfunction of a module-specific voltage sensor 5_1, 5_2, ..., 5_N can also be detected if its measured values ​​vary and no change is made to the system configuration state (and, if applicable, if the voltages recorded by the system-specific voltage sensor 4 show no changes).

[0093] Additionally, the following cases can be taken into account: If the state of a module M1, M2, ..., MN changes from activated to deactivated or vice versa, and the voltage jump observed by the system-specific voltage sensor 4 briefly, in particular within a period of up to 10 ms, ideally up to 10 µs, before and after the switchover does not correspond to the voltage jump detected by the corresponding module-specific voltage sensor, then a malfunction can be detected, in particular provided that no malfunction of the system-specific voltage sensor 4 has been detected.

[0094] Fig. Figure 4 shows a schematic flowchart of a method according to the invention for determining a distortion of the system-specific voltage sensor 4 in an open-circuit state. In this embodiment, the concept of phase-shifted counters (also known as phase-shifted carrier modulation) and sequential sampling is implemented, which has the advantage that all possibilities can be traversed for a system output voltage level without having to change the normal modulation strategy, and even distributed modulation strategies can be used.

[0095] In particular, the control and evaluation unit 9 sets the voltage level of a reference signal R to a system output voltage level-specific value, in particular to the value (j-0.5) / N, where j denotes the system output voltage level to be provided. A carrier signal generation unit 11 generates carrier signals T that are phase-shifted relative to each other. The phase of the carrier signals T is preferably uniformly distributed over one period of a carrier signal T. A carrier signal T is generated for each module M1, M2, ..., MN. The reference and carrier signals R, T are compared by a comparator 12. If the module-specific carrier signal T becomes greater than the reference signal R, the corresponding module M1, M2, ..., MN is activated. If it becomes smaller, the corresponding module M1, M2, ..., MN is deactivated. The control and evaluation unit 9, the carrier signal generation unit 11, and the comparator 12 can be part of the system described in Fig. The system 1 shown in Figure 1, in particular part of a modulation subsystem, can be used to set module states in a method for determining a distortion of the system-specific voltage sensor 4 as follows: i. After a start step S1, based on a predefined system output voltage level j, in a second step S2 all permutations of enabled and disabled module states, i.e., all possible system configuration states, that provide the desired system output voltage level j are determined. Then, a first system configuration state from this set is set. ii. In a third step S3, the module states are set depending on the output signal of comparator 12, which is provided in the form of switching pulses. This sets the various permutations and thus system configuration states. In each system configuration state, the system output voltage and the module voltage of the activated modules M1, M2, ..., MN are measured. After the third step S3, the procedure returns to the second step and the next permutation, i.e., the next system configuration state, is set. This continues until all system configuration states have been set. iii. In a fourth step S4, as explained above, the distortion of the system-specific voltage sensor 4 and, if applicable, the distortions of the module-specific voltage sensors 5_1, 5_2, ..., 5_N are determined. iv. In a fifth step S5, the distortion correction function can be updated. v. In a sixth step S6, the system output voltage level j is set to a changed value, in particular to the next higher system output voltage level.

[0096] The procedure can be performed by a monitoring subsystem of the in Fig. The system shown in 1 will be carried out.

[0097] The sampling times of the voltage measurement in the third step S3 are preferably synchronized with the rising or falling edges of the switching pulses generated by the comparator 12. For example, with carrier signals T having a triangular waveform, sampling T_C / 4N seconds after the detection of a positive edge can occur, where T_C is the reciprocal of the switching frequency of a module M1, M2, ..., MN.

[0098] Alternatively, the sampling time can be determined based on a counter value that defines the value of a carrier signal T. This value is incremented from a minimum to a maximum value and then decremented back to the minimum value or reset to the minimum value. For example, sampling can occur after a predetermined time interval following the reaching of a predetermined counter value, such as a maximum counter value. The reciprocal of the switching frequency T_C should be at least large enough to allow the measured values ​​and the modular electrical system 1 to reach a stable state. Ideally, it should be 10 times higher than a value dependent on the sensor / system dynamics, preferably 10N times higher (where N is the number of modules), and most preferably 100N times higher.

[0099] Fig. Figure 5 shows a schematic flowchart of a method according to the invention for determining a distortion of the system-specific voltage sensor 4 under a load condition. The method is essentially the same as that described in Figure 5. Fig. The procedures described in section 4 are described, and reference can be made to the corresponding explanations. In contrast to the procedure described in [reference to another document]. Fig. In the method described in Figure 4, however, the reference signal R is controlled by the control and evaluation unit 9, in particular depending on the load-related system output voltage level. Specifically, not all system output voltage levels can be traversed. Rather, the modulation sets all possible system configuration states to provide a load-related system output voltage level j and the next lower system output voltage level j-1.

[0100] In an intermediate step SZ, an unfiltered voltage signal can be determined based on the filter characteristics of the filters present in the modular electrical system and the measured, filtered voltage signal. In other words, an unfiltered voltage signal can be determined or reconstructed based on the filtered and subsequently acquired voltage signal, or a deviation between the unfiltered and the reconstructed voltage signal can be minimized. This method can advantageously compensate for, for example, the low-pass behavior of a voltage sensor. This is particularly beneficial when it is not possible to set the duration of a system configuration state long enough to prevent the sensor's low-pass behavior from affecting the voltage measurement.If the function f(x) describes the ideal output signal of the sensor (dynamic behavior) in response to an input signal x, but the voltage sensor, for example, exhibits slight delays and low-pass behavior, then the input signal can be determined using the relationship x = f^{-1}(f(x))).

[0101] If all modules M1, M2, ..., MN have the same or similar modulation indices, the system output voltage exhibits a pulsating waveform, particularly a square wave, with N times the switching frequency of a module M1, M2, ..., MN, with the j-th system output voltage level as the maximum value and the j-1-th system output voltage level as the minimum value. Here, the modulation index can denote the level of the reference signal used for comparison with the carrier signal. Each module can have a similar or an individual modulation index.

[0102] Again, the sampling times can be synchronized with the rising or falling edges of the switching pulses generated by comparator 12; for example, a sampling D·T_C / 2N can be performed one second after a rising edge to detect the j-th system output voltage level and (1-D)·T_C / 2N seconds after a falling edge. Here, D denotes a duty cycle.

[0103] Alternatively, as explained above, the sampling time can be determined based on a counter value that defines the value of a carrier signal T. For example, sampling can occur after a predetermined time interval following the reaching of a predetermined counter value, such as a maximum counter value. Furthermore, sampling to detect the j-th system output voltage level can occur immediately upon reaching the maximum value, and sampling to detect the j-1-th system output voltage level can occur immediately upon reaching the minimum value. The dependence on the counter value can be configured according to the shape of the carrier signal.

[0104] If the distortions of the module-specific voltage sensors 5_1, 5_2, ..., 5_N are also determined and corrected, this should preferably be done on the basis of measured values ​​determined with an updated sensor parameter of the system-specific voltage sensor 4.

[0105] Fig. Figure 6 shows a schematic flowchart of a further embodiment of a method according to the invention for determining a distortion of the system-specific voltage sensor 4 under a load condition. The method is essentially the same as that described in Figure 6. Fig. The procedures described in section 5 are described below, and reference can be made to the corresponding explanations.

[0106] While at the in Fig. In the embodiment shown in section 5, the system configuration states are set in a fixed sequence; this is done in the Fig. In the embodiment shown in Figure 6, the phase shift between the carrier signals T and / or the connection sequence of the modules M1, M2, ... , MN activated in the current system configuration state is changed, in particular after a predetermined time period, e.g. after a multiple of the period of a carrier signal or a multiple of a fundamental frequency of the system output voltage.

[0107] For this purpose, system 1 includes a counter 13 for determining the predetermined time duration and a device 14 for setting the connection sequence and / or the phase offset.

[0108] If, for example, the connection sequence of the module connection is determined by the vector P, where the first element is the ordinal number of the module M1, M2, ..., MN that is connected first, the second element is the ordinal number of the module M1, M2, ..., MN that is connected second, etc., then one possible variation of the sequence can be [1, 2, 3]⇒[2, 3, 1]⇒[3, 1, 2]. Similarly, the phase shifts of the carrier signals can also vary [ϕ_1,ϕ_2,ϕ_3 ]⇒[ϕ_2,ϕ_3,ϕ_1]⇒[ϕ_3,ϕ_1,ϕ_2]. This can be taken into account accordingly by the carrier signal generation device 11.

[0109] Fig. Figure 7 shows probability distribution functions of a system-specific voltage sensor 4 and a set of module-specific voltage sensors 5_1, 5_2, ... 5_N (see Fig. 2) A dashed line represents the probability distribution function of the system-specific voltage sensor 4. It indicates a static distribution of measured values ​​when exactly one voltage value is repeatedly measured and thus provides a measure of the measurement quality. The probability distribution function of the entire set of module-specific voltage sensors 5_1, 5_2, ... 5_N is determined based on the probability distribution functions of the individual module-specific voltage sensors 5_1, 5_2, ..., 5N, each of which indicates a static distribution of measured values ​​when exactly one voltage value is repeatedly measured by the corresponding voltage sensor.

[0110] The preceding explanations assumed that the module-specific voltage sensors 5_1, 5_2, ..., 5_1 are relatively uniform and exhibit similar tolerances (similar measurement accuracy). If this is not the case, a correction of the sensor distortions can be performed based on results from heterogeneous sensors, i.e., with different tolerances.

[0111] To account for sensor tolerances when determining distortion, the reciprocal of the relevant tolerances can be included in the distortion calculation. This approach advantageously ensures that measurements with smaller tolerances, and thus higher measurement accuracy, have a greater influence on the distortion determination.

[0112] For example, distortions can be determined taking sensor tolerances into account by using the reciprocal of the squared standard deviation in the equation for determining the distortion. For Gaussian-distributed sensor tolerances, such as those found in... Fig. As shown in Figure 7, the distortion can be calculated as follows: Bias_i=k_i(V_SLi−∑_(j=1 to n)V_MLij) where • V_SLi denotes the measured system output voltage at setting step n and the set i-th system configuration state, • V_MLij denotes the module voltage for module j in the nth step and the set i-th system configuration state, where n denotes the number of different system output voltage levels, • k_i is a confidence factor defined as the quotient of an area under the probability distribution function of the entirety of the activated module-specific voltage sensors 5_1, 5_2, ... 5_N and an area under the probability distribution function of the system-specific voltage sensor 4 in the area between the intersection points of the probability distribution functions.

[0113] After the bias for all permutations / system configuration states has been calculated in each step n, the average bias per step n is determined as follows. Bias_(avg,n)=(∑_i Bias_i) / (∑_i k_i)

[0114] By using this distortion equation, the effects of sensor tolerances can be effectively incorporated into the bias correction process, leading to a more accurate and reliable determination of the distortion. Reference symbol list 1 Sensor monitoring system 2 Energy source submodule 3 Switching submodule 4 Voltage sensor 5_1, 5_2, ..., 5_N voltage sensors 6 Sensor 7_1, 7_2, ..., 7_N Current sensors 8 Signal or data line 9 Control and evaluation unit 10_1, 10_2, 10_N Evaluation units 11 Carrier signal generation device 12 Comparator 13 counters 14. Setup for adjusting the connection sequence M1, M2, ..., MN Module N number of modules R Reference signal S1-S6 steps T carrier signal T1, T2 output connections

Claims

Method for sensor monitoring in a modular electrical system (1) comprising several modules (M1, M2, ..., MN) each with a source submodule (2) and a switching submodule (3) for activating or deactivating the module (M1, M2, ..., MN), comprising the steps: a) setting at least one system configuration state, b) acquiring a state-specific system voltage (V_o) with a system-specific voltage sensor (4) for the at least one system configuration state, c) determining at least one sensor parameter of the system-specific voltage sensor (4) and / or detecting a malfunction of the system-specific voltage sensor (4) as a function of the state-specific system voltage (V_o). Method according to claim 1, characterized in that a module-specific voltage sensor (5_1, 5_2, ..., 5_N) additionally detects a module voltage (V_j) of at least one module (M1, M2, ... MN), wherein the at least one sensor parameter is additionally determined as a function of the at least one module voltage (V_j). Method according to claim 2, characterized in that a sensor parameter of the module-specific voltage sensor (5_1, 5_2, ..., 5_N) is determined as a function of the state-specific system voltage (V_o) and the module voltage (V_j). Method according to one of the preceding claims, characterized in that a plurality of system configuration states are set and the corresponding state-specific system voltages (V_o) are detected, wherein the at least one sensor parameter is determined as a function of the detected system voltages (V_o). Method according to one of the preceding claims, characterized in that a) a first system configuration state is set in which all modules (M1, M2, ... MN) of the system (1) are deactivated, or b) a second system configuration state is set in which exactly one module (M1, M2, ... MN) of the system (1) is activated, or c) a third system configuration state is set in which exactly a number n of modules (M1, M2, ... MN) of the system (1) are activated, wherein the number n is greater than one. The method according to claim 5, characterized in that a state-specific system voltage (V_o) and a module voltage (V_j) of the respective activated module (M1, M2, ... MN) are determined for at least one second system configuration state, wherein the at least one sensor parameter of the system-specific voltage sensor (4) is determined as an updated sensor parameter depending on this state-specific system voltage (V_o) and the module voltage (V_j), wherein the module voltage (V_j) of the respective activated module (M1, M2, ...) is then determined for several different second system configuration states.MN) and the state-specific system voltage (V_o) is determined, wherein the state-specific system voltage (V_o) is determined with the updated sensor parameter, wherein the at least one sensor parameter of the module-specific voltage sensor (4) is determined as a function of the respective module voltage (V_j) and the state-specific system voltage (V_o). Method according to one of the preceding claims, characterized in that the determination of at least one sensor parameter is carried out depending on at least one sensor tolerance. Method according to one of the preceding claims, characterized in that the method is carried out in an idle state or in a load state of the system (1). Method according to one of the preceding claims, characterized in that the modules (M1, M2, ... MN) are activated at different times to set different system configuration states. System for sensor monitoring in a modular electrical system (1), comprising the modular electrical system (1), a device for setting at least one system configuration state, a system-specific voltage sensor (4) and at least one control and evaluation device (9), wherein the system (1) is configured to perform a method according to any one of claims 1 to 9.

Images