DE102022001302A1 - HD electrical inverter system and associated control method - Google Patents

Abstract

It is known that cascaded H-bridges with heterogeneous module voltages can theoretically produce exponentially refined output voltage levels and achieve high output resolutions with only a single power source. However, several critical aspects remain to be thoroughly investigated, namely 1) the controllability of module voltages under arbitrary load conditions and 2) the optimal scheduling of switching states.We present a scheduling algorithm and topological changes for binary asymmetric cascaded H-bridges (ACHBs) to achieve a to achieve completely sensorless operation under all load conditions while using only one power source. The scheduling algorithm cancels out the energy consumption of the floating modules per time frame, while low-power DC auxiliary circuits, which naturally arise from the bridge structure, can correct for any voltage drift without the need for voltage sensors. Rigorous analyzes reveal an inherent performance trade-off between reference sampling accuracy and floating module voltage controllability for ACHBs in general, for which the proposed algorithm always achieves the Pareto front. The proposed ACHB and algorithm are tested in a six-module experimental setup with 65 output levels (i.e., effectively 6 bits).

Description

The invention relates to a circuit topology and an associated control method for a power electronic inverter with ultra-high output voltage resolution, high efficiency and high scalability to cover small, medium and high powers, but with low costs and operating effort.

The need for high power power sources arises from many new applications that often simultaneously require high power (> 1 kW, sometimes > 100 kW or even > 1 MW), high quality (e.g. resolution of > 10 bits) and a require high frequency (> 1 kHz). Examples include magnetic nerve stimulation devices, high-performance audio amplifiers, and high-performance electric motor test benches. However, most existing power source technologies have been developed and refined with a specific operating range in mind, and there is no general solution that can deliver high-resolution, flexible AC waveforms from kW to MW and frequencies from DC to MHz. Customization has always been the main approach to power converter development, which requires large upfront investments in each new R&D project and delays market entry. There is also a need for fast, flexible hardware prototyping in academia, with each lab building another specific and expensive power converter setup.

Scientific reference B. Wang et al. (2022) [Wang, Boshuo, Zhongxi Li, Charles E. Sebesta, Daniel Torres Hinojosa, Qingbo Zhang, Jacob Robinson, Gang Bao, Angel V. Peterchev, and Stefan Goetz (2022). “Multichannel power electronics and magnetic nanoparticles for selective thermal magnetogenetics”. Journal of Neural Engineering, doi: 10.1088/1741-2552/ac5b94] describes a circuit that can generate high-frequency power with high quality and performance. However, the solution described there can only generate very specific frequencies, but is not able to generate practically any power within a specific bandwidth.

Z. Li et al. (2019) [Li, Zhongxi, Ricardo Lizana, Zhujun Yu, Sha Sha, Angel V. Peterchev, and Stefan M. Goetz (2019). “A modular multilevel series/parallel converter for a wide frequency range operation.” IEEE Transactions on Power Electronics, 34(10): 9854-9865. doi: 10.1109/TPEL.2019.2891052], S. Goetz et al. (2016) [Goetz, Stefan M., Zhongxi Li, Xinyu Liang, Chengduo Zhang, Srdjan M. Lukic, and Angel V. Peterchev (2016). “Control of modular multilevel converter with parallel connectivity-Application to battery systems.” IEEE Transactions on Power Electronics, 32(11):8381-8392, doi: 10.1109/TPEL.2016.2645884], patent specification US 9,502,960 and patent specification US 9,496,799 describe modular power electronic circuits that can produce a wide range of output signals with high power, but the quality of the output is low and requires a large number of modules for high-quality output voltage, e.g. B. with a resolution of > 10 bits.

Z. Li et al. (2022) [Li, Zhongxi, Jinshui Zhang, Angel Peterchev, and Stefan Goetz (2022). Modular Pulse Synthesizer for [Transcranial Magnetic Stimulation with Flexible User-Defined Pulse Shaping and Rapidly Changing Pulse in Sequences. arXiv preprint arXiv:2202.06530.], Z. Zeng et al. (2022) [Zeng, Zhiyong, Lari Koponen, Rena Hamdan, Zhongxi Li, Stefan Goetz, , and Angel V. Peterchev (2022). “Modular multi-level TMS device with wide output range and ultrabrief pulse capability for sound reduction”. Journal of Neural Engineering, 19(2): 026008, doi: 10.1088/1741-2552/ac572c], and S. Goetz et al. (2012) [Goetz, Stefan M., Michael Pfaeffl, Jonas Huber, Matthias Singer, Rainer Marquardt, and Thomas Weyh (2012). “Circuit topology and control principle for a first magnetic stimulator with fully controllable waveform”. In 2012'Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 4700-4703. doi: 10.1109/EMBC.2012.6347016] disclose devices that can generate virtually any form of output with very high power and wide bandwidth. However, the output quality is rather limited and requires many components or modules since the voltage granularity only grows proportionally to the number of modules.

Scientific Reference J. Fang (2021) [Fang, Jingyang, Frede GE Blaabjerg, Steven Liu, and Stefan Goetz (2021). “A review of multilevel converters with parallel connectivity”. IEEE Transactions on Power Electronics, 36(11):12468-12489. doi: 10.1109/TPEL.2021.3075211] discloses various module configurations, some of which may be used in combination with the invention and are intended to be incorporated herein by reference.

The patent specification US 5,963,086 describes an amplifier circuit that can produce a wideband output with high power and efficiency, but the output quality is so low that a filter is required to eliminate some of the distortion, which comes at the expense of efficiency and reduces the possible bandwidth and dynamics. In addition, the bandwidth is limited to the fastest switching speed of the transistors. It is not possible to exceed this speed by distributing the switching processes over several transistors.

The amplifier circuit of the patent specification US 8,149,061 On the other hand, it can reach very high frequencies, but the bandwidth and efficiency are limited.

The device from the patent specification US 8,604,883 can produce virtually any power of significantly higher quality than the above solutions, but the efficiency is very low, so higher powers cannot be produced without unacceptable power loss or overheating of the amplifier.

The most common method for achieving high performance while maintaining high output quality are so-called cascaded H-bridge converters (CHB). CHB is a typical modular topology, which primarily consists of several identical power switching modules. By turning these modules on and off individually, the CHB generates a range of discrete voltage levels. Since the number of output voltage levels grows in proportion to the number of modules, a large number of modules are required for high-resolution output. For example, a 7-bit output resolution requires an order of 128 modules; further increasing the resolution to 8 bits doubles the number of modules. The enormous component cost and operational overhead hinder the use of CHBs in low/medium voltage applications, despite their excellent scalability in terms of power and output resolution.

Digital circuits such as Sigma/Delta DAC, on the other hand, easily output 16 bits in a very wide bandwidth. However, such circuits can hardly carry any meaningful electrical load, which makes them unusable in power electronics. General amplifier circuits that are characterized by lower power can produce the quality but not the power because their efficiency is limited. Most amplifiers use transistors that operate linearly in the current or voltage controlled range, resulting in significant losses.

The invention described below introduces a kind of asymmetric cascaded H-bridge circuit (ACHB) that solves the above problems and forms a high-definition converter (HD converter). The proposed ACHBs retain the excellent (performance) scalability of CHBs while possessing the high resolution and bandwidth of sigma/delta DACs. For example, the proposed ACHB requires only 7 modules for 7-bit resolution and 8 modules for 8-bit resolution - a significant simplification compared to traditional CHBs.

The ACHB can have a modular design so that it includes several circuit modules. However, the ACHBs intentionally control the modules at different voltage levels. The module voltages can e.g. B. form a power sequence of {2 ^k }, which enables sigma/delta synthesis. In addition to using a mix of different module voltages for high resolution output, the invention discloses a solution to how such a circuit would perform under arbitrary load conditions. The combination of control method and circuit implementation prevents ACHB modules from absorbing unequal amounts of energy, deviating from their specified voltages and causing the system to quickly collapse.

The present invention is not subject to any compromise between the output load range and the need to implement a large number of galvanically isolated DC supplies (e.g. one per module): the proposed ACHB can even operate with only one DC supply while accepting all types of load conditions .

Summary of the invention

An important embodiment of the invention includes a series of DC/AC circuit modules, each charged with a specific voltage. Adjacent modules are electrically connected to each other, e.g. B. by a power cable and a small inductance (e.g. 0.1 µH to 10 mH, preferably 1 µH to 1 mH). The coil can be e.g. B. a connection with multiple turns, a connection with magnetic material that forms a closed volume, or a single coil component. The positions of the circuit modules are sorted by voltage. One The preferred embodiment of the invention includes N bridge-like circuit modules charged to different voltages, e.g. B. proportional to the number sequence 2: 4: 8: ...: 2 ^N.

The invention is presented below using four hierarchy levels: module, inter-module connection, arm formation and control.

module

The invention can in principle be combined with a large number of module circuits (so-called microtopologies or module topologies) of modular converters. Preferred module types are two-quadrant modules with pure series connection (hereinafter referred to as M2C-2q), four-quadrant modules with pure series connection (hereinafter referred to as M2C-4q), short-circuit-proof modules with pure series connection (hereinafter referred to as 4q-KGM2C). , four-quadrant modules with series and parallel connection (hereinafter referred to as M2SPC-4q), two-quadrant modules with series and parallel connection (hereinafter referred to as M2SPC-2q). These modules work with different voltage levels, which preferably differ from each other by a factor of 2 ^k or 3 ^k , ie with voltages that, to a first approximation, follow a power series.

Inter-module connection

The inter-module connection or connection for short refers to the electrical connection between adjacent modules. The adjacent modules are connected to each other via an electrical conductor or conductor and a choke coil, inductor or inductor. While the electrical conductor acts as a so-called power path, such as. B. a power cable or busbars - similar to conventional CHBs - primarily conducts the load current, the additional inductor offers a possibility for controlled energy transfer across the connection. The power cable, the inductance and the energy storage elements, e.g. B. capacitors, battery cells or batteries, the two adjacent modules form a bidirectional DC-DC converter for energy transmission. Such energy transfer is crucial for maintaining the exponential voltage gradient between modules.

poor

A series of modules connected together is called an arm. Multiple arms can be connected together to form different macro topologies for different needs. For example, three of these arms can be connected together to form a so-called star configuration for three-phase applications such as motor drives. A connection can be made with electrical connections, such as cables and wires, and inductors for maximum controllability ( 2 B) can be realized, effectively sharing the connection between multiple modules from different arms. An alternative and more advantageous approach is to merge the end modules of the arms at their DC connections and remove the unnecessary transistors.

steering

The control method presented uses a continuous reference waveform as an input signal and determines the switching states, also briefly states, of electrical switches, e.g. B. power transistors to approximate the reference waveform with a predetermined sampling frequency. Importantly, the presented method 1) guarantees a narrow output error range, strictly limited by the DC link voltage of the smallest module, 2) guarantees a zero net charge for all floating modules, resulting in significantly simplified voltage regulation of all DC links modules, and 3) exploits the redundant switching states of the neighboring modules and the connection to ensure the specified voltage ratio without the need for voltage sensors.

Floating in the sense of the invention means in particular that the elements, for example energy storage devices or modules, do not have a fixed potential reference to earth but can be shifted in their potential relative to earth, for example by effectively activating switching elements.

The above features, together with the above-mentioned converter topologies, enable highly accurate, preferably voltage-fed, waveform synthesis under virtually any load conditions, whereby the predefined DC link voltage ladder is maintained. Only a DC voltage source is required and a sensor is not required. The computational complexity is on the order of O(N), where N is the number of modules per branch.

The required properties are made possible by innovations in the control system and circuit topology. At its core, the control method divides the sampled reference waveform into segments of a specific length (typically on the order of 100 µs). For each segment, all samples are considered together in order to exploit switching redundancy. This is the core idea to achieve low output error rate and high DC stability at the same time. In contrast, scoping the control for each waveform sample - as in most existing control methods for ACHBs - either fails to find the best compromise between the aforementioned control objectives or requires the transfer of information to the future control cycles, complicating the stability of the system and potentially lead to unpredictable behavior. Although batching introduces throughput latency, this latency is on the order of the switching speed (e.g., ten times the switching period of the fastest transistor in the ACHB system), which is negligible in most applications. Thanks to the high vertical resolution of the proposed converter, large filters are not required, so the effective throughput latency is expected to be even shorter than traditional two-/three-stage converters aiming at the same waveform quality.

Short description of the characters

• 1 zeigt schematische Darstellungen von drei elektrischen Umrichtersystemen nach dem Stand der Technik
  • 1a ( US 8,441,224 B2 ) zeigt ein häufig verwendetes elektrisches Umrichtersystem für Motorantriebe, bei dem eine fest verdrahtete Batteriebank eine feste Gleichspannung liefert, die einen DC/AC-Wechselrichter speist und schließlich die Last antreibt.
  • 1b ( US 7,279,855 B2 ) zeigt ein elektrisches Umrichtersystem, das der Erfindung in 1a ähnelt, bei dem jedoch eine DC/DC-Stufe zwischen der Batteriebank und dem DC/AC-Wechselrichter eingefügt ist.
  • 1c-d ( US 8,760,122 B2 , US 20190288617 A1 ) zeigen ein elektrisches Umrichtersystem, das hauptsächlich aus elektrischen Modulen besteht. Die Module bilden Strings, die direkt die elektrische Last speisen.
• 2a-c zeigen verschiedene Ausführungsformen der Erfindung.
  • 2a ist für einphasige Anwendungen gedacht.
  • 2b ist für dreiphasige Anwendungen mit einer Mindestanzahl von Modulen gedacht. Die Modulstränge sind sternförmig angeschlossen.
  • 2c ist für dreiphasige Anwendungen gedacht, die sich durch eine minimale Anzahl von Transistoren pro Modul und eine bessere Steuerbarkeit auszeichnen.
• 3a-c zeigen verschiedene Ausführungsformen der Module mit unterschiedlicher Schaltredundanz, Polarität und Anzahl der Komponenten.
  • 3d zeigt eine Ausführungsform des zentralen Schaltmoduls, das verschiedene Konverterarme miteinander verbindet.
• 4a-b zeigen verschiedene Ausführungsformen der Zusammenschaltung, die eine eigene Gleichstromwandlung zwischen den Modulen ermöglicht.
• 5a-b zeigen die grundlegenden Schaltzustände. Die Schaltzustände sind durch den Leistungspfad und den Induktorpfad (Gleichspannungswandler) entkoppelt.
• 6 zeigt Ausführungsformen des erfindungsgemäßen Steuerungsverfahrens.
• 7 zeigt eine Ausführungsform eines erfindungsgemäßen Kontrollverfahrens.

Further advantages and further developments of the invention can be found in the description and the figures.

• 1 shows schematic representations of three electrical converter systems according to the state of the art
  • 1a ( US 8,441,224 B2 ) shows a commonly used motor drive electrical inverter system in which a hard-wired battery bank provides a fixed DC voltage that feeds a DC/AC inverter and ultimately drives the load.
  • 1b ( US 7,279,855 B2 ) shows an electrical converter system according to the invention 1a similar, but in which a DC/DC stage is inserted between the battery bank and the DC/AC inverter.
  • 1c-d ( US 8,760,122 B2 , US 20190288617 A1 ) show an electrical inverter system consisting mainly of electrical modules. The modules form strings that directly power the electrical load.
• 2a-c show various embodiments of the invention.
  • 2a is intended for single-phase applications.
  • 2 B is intended for three-phase applications with a minimum number of modules. The module strings are connected in a star shape.
  • 2c is intended for three-phase applications characterized by a minimum number of transistors per module and better controllability.
• 3a-c show different embodiments of the modules with different switching redundancy, polarity and number of components.
  • 3d shows an embodiment of the central switching module that connects different converter arms to one another.
• 4a-b show various embodiments of the interconnection, which enables its own direct current conversion between the modules.
• 5a-b show the basic switching states. The switching states are decoupled by the power path and the inductor path (DC-DC converter).
• 6 shows embodiments of the control method according to the invention.
• 7 shows an embodiment of a control method according to the invention.

Detailed description of the invention

contraption

2a shows an embodiment of the high resolution converter (herein referred to as HD converter or HS converter) 210 for single-phase applications. This HD converter 210 consists of a large number of topologically identical units 300 that are connected in series. Each unit 300 consists of a switching module 310 and an interconnect 400. The switching modules 310 can be controlled to feed multiple discrete voltages into the HD converter 210 and ultimately provide power to the electrical load 211. Within the units 300, the connection part is implemented to provide an interface for a variety of switching modules 310. The connection is made with at least one conductor and at least one coil. While the conductor or conductors conduct the majority of the load current through the HD converter 210, the coil or coils enable controllable energy transfer between the switching modules 310 of the connected units 300. Therefore, the choke coil of the connection 400 can be used, to regulate the excitation level of the electrical modules 310. Importantly, the electrical modules 310 can be energized with a list of different voltage values, e.g. B. with a base 2 exponential number series (i.e. 1, 2, 4, ..., ^2N ) to maximize the total number of voltage levels output. The choice of the number of units 300 depends primarily on the desired number of output voltage levels (ie, on the desired output quality of the HD converter), while the maximum output voltage of the HD converter 210 is primarily proportional to the largest unit 300, which by definition, the switching module 310 includes the highest excitation voltage. The maximum current of the HD converter 210 is defined by the minimum current rating of all units 300, but excluding any components directly connected to the inductor since the inductor is only involved in energizing the units 300 and not the load 211 contributes.

2 B shows an embodiment of the HD converter 220 suitable for three-phase applications. The three-phase HD converter 220 can be obtained by combining three single-phase HD converters 210 with appropriate adjustment on the central switching module 320. Similar to the single-phase HD converter 210, the devices 300 are each operated so that the internal switching module 310 is energized at a certain voltage level. The difference is that the target voltage levels can be even further apart, preferably forming an exponential number series to base 3 (i.e. 1, 3, 9, ..., ^3N ) to produce an even larger number of output voltages from the same number of To achieve switching modules 310 per phase.

2c shows another version of the HD converter 230, which is suitable for three-phase applications with higher maximum voltage. The three-phase HD converter 230 can be obtained by assembling six HD converters 210, which form three AC terminals 232 and a combined DC bus 234. Power can flow in either direction between the AC ports 232 and the combined DC bus 234. Although aimed at similar applications, the HD Converter 230 provides better control redundancy than the previous three-phase HD Converter 220 design.

The operation of the HD converters 210/220/230 depends crucially on the modules remaining at their intended voltage levels. We refer to this effort to control the module voltage as voltage regulation. At its core, the voltage regulation adds or removes a certain amount of charge from the energy storage 313 - often in the form of capacitors - so that the energy storage remains exactly at the intended voltage level in order to deliver the exact voltage level when activated by the HD converter . Voltage regulation can be achieved through various approaches, the simplest of which is to connect galvanically isolated bi-directional power supplies to the energy storage 313 of each module. Of course, such an approach significantly increases the cost and energy loss of the HD converter system. In this invention, voltage regulation is achieved via the load current itself and the embedded DC interface. At a higher level, the proposed control method activates the switching modules 310 so that the net change in charge of the energy storage 313 of all modules 310 is almost zero, while the embedded DC interface compensates for the remaining energy deficit, if there is one. The specific circuit topologies that enable such operation are presented below, followed by the description of the control method.

3a-c show various embodiments of the switching module 310. The switching module 310 contains a plurality of power switches 315 and at least one energy storage 313. Each switching module has at least four connections, of which at least two are used to conduct the main load current of the HD converter 210/220/230 and the remaining connections are connected to the inductances of the connections 400. When characterizing switching modules 310a-c, it is common practice to connect two power switches 315 in series and thus form a so-called half-bridge 311 ( 3a) . The upper connection 314a/d and the lower connection 314c/f of a half bridge 311 are connected to the energy storage 313. The middle connection 314b/e supplies the output of the half bridge 311. The two power switches 315 of the half bridges are normally switched on and off in a complementary manner.

3a shows an embodiment of the switching module 310a. Here, the combinations of ON and OFF states of the switches 315 effectively connect the at least one energy storage 313 to the main terminals 314b, 314e in different polarities. Each of these temporal switching combinations is referred to as a switching state. The module variant 310a has four circuit breakers 315 or two half bridges 311. Looking at the two main connections 314b and 314e, there are four switching states: the positive state 310a1, the negative state 310a2 and two bypass states 310a3 and 310a4. The naming of the switching states reflects the polarity with which the energy storage 313 is inserted into the HD converter. The remaining auxiliary terminals 314a/c/d/f do not have to share the load current, but rather serve to form a DC-DC converter interface with the adjacent interconnection 400. In particular, due to the inductance of the interconnection 400, which shields the voltages of the energy storage 313, it is possible to use such a DC voltage interface to achieve a desired intermediate circuit voltage ratio via the interconnection and thus all intermediate circuit voltages of the switching modules 310 to any number of rows steer. The switching module 310a is the simplest circuit that allows simultaneous control of the output voltage and the DC interface. However, the two control targets are not completely decoupled since the auxiliary terminals 310a/c/d/f share the same set of half bridges 311 with the main terminals 310b/e. Often the regulation of the output voltage takes priority, which affects the regulation of the DC link voltage. To complement this hardware shortage, the control algorithm can coordinate all units 300 at the HD converter level to provide additional redundancy. The control algorithm is described in detail below.

3b shows a variation of 3a , which has eight circuit breakers 315 or four half bridges 311. By adding more power switches 315, two half bridges 311 can be used for the DC interface via corresponding auxiliary terminals 314b2 and 314e2. In this way, the control of the main output voltage and the DC interface are decoupled at two levels: at the control level, the switching states at the level of the HD converter do not need to be coordinated to resolve the conflict between the output quality and the regulation of the DC connection; At the circuit level, the power switches 315 intended for the DC interface may have a much lower current rating than the other power switches.

3c shows an embodiment of the switching module 310d. In this embodiment, the number of half bridges 311 is further increased to enable better controllability of the DC connections. Here, a pair of main terminals 314b1/e1 are used to conduct the main current, while the remaining terminals 314b2/b3/e2/e3 are connected to the inductances 400 of the connections.

3d shows the central switching module 320 of the HD converter 220. The central switching module 320 is where all the various arms are connected together. The central switching module includes the direct current source 312, the energy storage 313 and a plurality of half bridges. The half bridges are grouped evenly, from which different phases are branched off. The central switching module 320 is preferably designed as the module with the highest voltage with respect to each phase. In this way, the additional current required for voltage regulation between the different phases is minimal.

4a and 4b show four embodiments of the connections 401, 402, 403 and 404. Each of the connections includes at least one inductor and one conductor that complete the above-mentioned DC-DC interface between the switching modules 310a/b/c.

Connection 401 is suitable for four-quadrant operation, where the arm alone can supply voltage and current in any combination of polarities. The switching module 310a can also be configured into other connections 402 and 403. In this case the connections offset the adjacent switching modules 310a systematically so that the resulting arms can deliver either only non-negative or only non-positive voltages. The reduction in operating range (i.e. two fewer quadrants) comes at the cost of better controllability of the DC interface. These connections 402 and 403, in which the switching module 310a is involved, are particularly suitable for the three-phase HD converter 230.

In the interconnection 404, the more advanced switching module 310b is used instead of the switching module 310a to achieve better controllability of the DC-DC interface without compromising the operating range (i.e., maintaining the four-quadrant operating range). In this case, the four half-bridges 311 of the module 310b are intended for different tasks. The two half bridges 311, which are connected to the normal conductor of connection 404, conduct the full current of the HD converter systems. The half bridges connected to the choke, on the other hand, only conduct the current to complete the voltage regulation. If the HD converters 210/220/230 are controlled appropriately, most of the voltage regulation takes place via the load current, so that only a negligible part of the energy has to be transmitted via the DC voltage interface. Therefore, these half bridges 311, which are connected to the inductors of the interconnection 404, can preferably be designed with smaller rated currents in order to minimize costs.

steering

A control method according to the invention relates to a method that determines the switching combinations of the HD converter systems 210/220/230 in order to generate (stage I) the desired arbitrary output voltage form under any load conditions and (stage II) the voltage of the floating capacitors at the intended to maintain voltage series. No floating power supplies or voltage sensors are required.

5a shows the basic switching states of a generic switching module 310b. Here it is sufficient to focus only on the combination between the power path switching states 310b1 and the inductor path switching states 310b2. While the power path switching states 310b1 determine the output voltage level, the added current loop via the inductor path switching states 310b2 regulates the magnetization state of the inductor at junction 400. By alternating (de)magnetizing the inductor from the left and right modules, the method can generate electrical energy transmitted via the switching modules 310. In this way, the voltage state of the neighboring modules can be effectively controlled. 5b shows all 16 combinations between the switching states of the power path 310b1 and the inductor path 310b2 for one embodiment. These 16 switching combinations are categorized by how they change the system's output voltage (used by Stage I), and are further differentiated by how they magnetize the inductance of the two adjacent capacitors V _L (left) and V _R (right) (used by Stage II).

At any point in time, control stage I aims to create a column of states in 5b based on the requested output voltage, while the specific candidate within this column is to be determined by stage II for capacitor voltage regulation.

Stage I: Output voltage control

Stage I aims to synthesize the target output waveform with the highest possible quality while eliminating the accumulated charge on each floating module. This is a crucial step to ensure a (near) 0 energy deficit on the floating capacitors, i.e. sensorless voltage regulation. In a typical voltage series set by the number series ^2N , there is some redundancy in the synthesis of the output voltage that can be exploited by stage I. Unfortunately, this redundancy is not sufficient for certain voltage levels. A method according to the invention can therefore relax the output waveform requirements by allowing a transient output level error consistent with our finest module voltage. With 8 potential-free modules and therefore -256 to +256 possible output levels, the voltage curve is relaxed so that a transient error of -1 or +1 is permissible. Such a setting results in negligible distortion in most practical scenarios, but increases redundancy to achieve a 0 energy deficit for all modules.

6 shows the overall architecture of the control. The input buffer 701 accumulates the stream of target output voltage levels into sections of a predetermined length L. Each section of the waveform is then processed by the Stage I (Stage I) control algorithm 702. The output switching vectors only indicate the desired polarity of the power path switching state 310b1, which is also one of the columns of the switching state lookup table (LUT) in 5b corresponds. The specific selection within each selected column is also determined by the capacitor voltage correction stage (stage II) 704.

As previously mentioned, the output voltage control (Stage I) 702 must approximate the specified output reference section with the accuracy of the finest voltage level while ensuring a zero energy deficit on all floating modules. For this purpose, we propose the following algorithm for 702.

Very importantly, the algorithm has the following features: (1) The total absolute output voltage error per section is limited by ^2N , regardless of the frame length; in other words, increasing the length of the input buffer 701 reduces the error density. By design, the maximum transient voltage error is limited by the smallest module. (2) The net input energy per floating module is zero assuming a nearly constant current per buffer length (i.e. in most practical cases). (3) The proposed algorithm achieves the Pareto front of (1) in terms of the trade-off between the buffer length and the output error density while satisfying (2). All of these properties are mathematically proven. Furthermore, (4) the proposed algorithm can be executed on the order of O(LN) iterations with appropriate parallelization of vector arithmetic. The time complexity O(LN) can be amortized as O(N) per time step over the course of the N-step waveform section. Considering that the number of output stages grows exponentially while the computation time increases only linearly, the proposed algorithm is feasible in mainstream FPGAs and has great scalability.

Stage II: Capacitor voltage control

After an energy deficit of almost 0 per floating module was already achieved in Stage I, it is the task of Stage II to further reduce the remaining net energy. This is done through exploitation tion of redundancy in each column of 5b . Independent of the switching states 5b the connection inductor is activated at any time $$L_{\text{int}}\frac{\text{d}{i}_L}{\text{d}t}=s_L{v}_L+s_R{v}_R,\left(s_L,s_R\in \left\{-\mathrm{1,}+1.0\right\}\right)$$

dem Gleichgewicht nähert, können wir jedes Spannungsverhältnis über $$\frac{V_R}{V_L}=-\frac{{\displaystyle \sum \tau s_L}}{{\displaystyle \sum \tau s_R}}$$

wobei τ die diskrete Regelkreiszeit ist, wobei V_R /V_L die Kondensatoren links/rechts von der Verbindungsinduktivität sind. Stufe I kann durch einen billigen Digitalzähler realisiert werden, der die Zählungen von S_L und S_R innerhalb eines vordefinierten Zeitfensters verfolgt. Es ist mathematisch bewiesen, dass bei Stufe I immer genügend Redundanz innerhalb von 5b vorhanden ist, um beliebige V_R /V_L zu erreichen. Insbesondere können wir V_R /V_L = 2: 1 an allen Verbindungsstellen einstellen, um die gewünschte exponentielle Spannungsreihe zu erreichen.Since the system ${\displaystyle \int L\frac{\text{d}{i}_L}{\text{d}t}}\text{d}t={\displaystyle \int s_L\left(t\right)v_L+s_R\left(t\right)v_R\text{d}t}=0$

approaching equilibrium, we can transcend any tension $$\frac{v_R}{v_L}=-\frac{{\displaystyle \sum \tau s_L}}{{\displaystyle \sum \tau s_R}}$$

where τ is the discrete control loop time, where V _R /V _L are the capacitors to the left/right of the connection inductance. Stage I can be implemented by a cheap digital counter that tracks the counts of S _L and S _R within a predefined time window. It is mathematically proven that at level I there is always enough redundancy within 5b is present to achieve any V _R /V _L. In particular, we can set V _R /V _L = 2:1 at all junctions to achieve the desired exponential voltage series.

Other embodiments

An embodiment of the invention comprises at least three modules, each of the modules comprising at least one energy storage element and at least three electrical switches, each of the modules being electrically connected to at least a second one of the modules by at least one electrical connection between the modules, and wherein the at least two electrical switches in each module are configured to change an electrical connection between a first module and at least a second module between at least two alternatives of electrical connection by effectively switching the electrical switches; and wherein at least two modules are connected by at least two electrical connections, at least one of which has inductance greater than 100 nH.

The at least one electrical connection with an inductance of 100 nH can be an inductor with magnetic material.

The at least two alternatives of the electrical connection can include at least one series state in which the at least one energy storage element of a first module is connected in series with at least one energy storage element of at least a second module.

In another embodiment, the at least one bypass state is in which at least one energy storage element of a first module is electrically connected to a second module with at most one of its electrical contacts.

A preferred embodiment comprises at least three branches, each with at least two modules, the three branches being connected by a paired electrical connection of at least one module of each branch. In this embodiment, at least a part of the paired electrical connection of at least one module of each branch can be connected to the energy storage element of at least one module of each branch so that it is connected to the positive terminals of the energy storage elements of the at least one module of each branch and at least another part of the pairwise electrical connection of at least one module of each branch to the negative connections of said energy storage elements of the at least one module of each branch is connected so that the energy storage elements are permanently connected in parallel or form a unit.

Upon change, the at least three electrical switches of a first module, along with the at least three electrical switches of a second module electrically connected to the first module, may dynamically transition between multiple connection states determined by a combination of multiple power path switching states and multiple inductor paths switching states are formed.

These multiple connection states can include at least four switching states of the power path and at least four switching states of the induction field, which can be freely combined by correspondingly switching the at least three electrical switches on and off.

For supply, a DC power supply, which does not have to be ungrounded or galvanically isolated, can be connected to the at least one energy storage element of at least one module, so that the DC power supply is electrically connected in parallel to the energy storage element. The DC power supply can e.g. B. be connected to ground with at least one of its connections.

• einen Leistungserzeugungsschritt, der Leistungszustände mehrerer Module derart bestimmt, dass die Verschaltung der jeweils auf ein individuelles Spannungsniveau aufgeladenen Energiespeicherelemente der entsprechenden Module entsprechend den gewählten Leistungspfad-Schaltzuständen eine Annäherung an die angeforderte Ausgangsspannung mit einem Fehler von höchstens der kleinsten Spannung eines der Energiemodulelemente der Module erzeugt; und einen Kondensatorspannungsladeschritt, der Induktorpfadzustände bestimmt und die ausgewählten Leistungszustände der Module verwendet, um für jedes von mindestens einem ersten und mindestens einem zweiten Modul einen Modulzustand auszuwählen, der den mindestens einen Induktor magnetisiert oder den mindestens einen Induktor, der das mindestens eine erste und das mindestens eine zweite Modul verbindet, entmagnetisiert, wobei das Magnetisieren des Leiters.

A corresponding control method for generating at least one powerful, high-quality electrical output voltage by the invention preferably carries out at least the following steps:

• a power generation step that determines the power states of several modules in such a way that the interconnection of the energy storage elements of the corresponding modules, each charged to an individual voltage level, generates an approximation to the requested output voltage with an error of at most the smallest voltage of one of the energy module elements of the modules in accordance with the selected power path switching states ; and a capacitor voltage charging step that determines inductor path states and uses the selected power states of the modules to select, for each of at least a first and at least a second module, a module state that magnetizes the at least one inductor or the at least one inductor that magnetizes the at least one first and the at least a second module connects, demagnetized, magnetizing the conductor.

Depending on the power path level, the same inductor path condition can have different effects. Whether an inductor path state is magnetizing, that is, charging an inductor that draws energy from the module energy storage elements, or demagnetizing, that is, discharging the inductor and charging the energy into a module energy storage element, depends on the power path state on the same module circuitry. A module connection (shown in 5a) is typically formed by the transistors, the electrical connection and the at least one inductor electrically located between the energy storage elements of at least two connected modules. The combination of power path state and inductor path state, also referred to here as the power inductor path combination, together form the overall module state, or module state for short or connection state. The magnetization of the inductance is increased when one of the charged energy storage elements, e.g. B. capacitors, of at least two modules that are connected by the inductor, a voltage is present or a series connection of the charged energy storage elements is effectively switched by a suitable magnetizing power-inductor path combination, so that the voltage on the at least one inductor in the same direction as the current goes through the inductor, so that the inductor becomes magnetically charged. The magnetization of the inductor is reduced when a voltage formed by one of the charged energy storage elements of at least two modules connected by the inductor or a series connection of the charged capacitors is effectively connected by an appropriate magnetizing power-inductor path combination, so that the voltage across the at least one inductor goes in the opposite direction as the current through the inductor, so that the inductor becomes magnetically charged. The magnetic energy of the at least one inductor is charged into the capacitors mentioned. The magnetization is not changed or practically not changed, ie maintained, when there is no or only a negligible voltage across the at least one inductor. A voltage that represents the voltage drop of the current through the coil across the internal resistances of the switches, such as transistors, electrical connections and the coil, can be neglected.

If, in alternately set magnetizing and demagnetizing inductor path states, different energy storage elements are connected in a circuit with at least one inductor, then in the magnetizing inductor path states, energy is transferred from the at least one first energy storage element, which is connected to the at least one inductor, to the at least one second energy storage element , which is electrically connected to the at least one inductor in the demagnetizing inductor path states, and a power transfer direction from the at least one first to the at least one second energy storage element can be defined.

The magnetization of the at least one inductor can be achieved in this embodiment by combining a bypassL power path state with an I+ inductor path state or a bypassH power path state with an I- inductor path state; and wherein the demagnetization of the at least one inductor is achieved by combining a bypassL power path state with an IH inductor path state or a bypassH power path state with an IH inductor path state, such that the magnetization energy of the inductor is effectively transferred into at least one electrical energy storage element.

The step of charging the capacitor voltage may further preferably maintain the magnetization of the at least one inductor, wherein the magnetization does not change substantially, and wherein maintaining the magnetization includes at least the combination of a series+ power path state with an I+ inductor path state or a series--power path state with an I--inductor path state or a BypassL power path state with an IL inductor path state or a BypassH power path state with an IH inductor path state.

The time duration of at least one magnetizing and demagnetizing inductor path state between at least a first and at least a second module can be selected so that it reflects the desired voltage ratio between the energy storage elements of the at least first and at least second modules. For example, the duration of the magnetization or demagnetization inductor path step can be selected mathematically depending on the voltage ratio of the two modules, e.g. B. at least within a range of stress ratios with an affine linear relationship of ax + b.

In addition, a preferred control method may also include a transfer state in which the selected states of the power path and the inductor path are transferred to the hardware.

Preferably, the capacitor voltage charging step further determines the inductor path states by receiving the power path states from the output generating step; calculates a voltage ratio for the required target voltage of the at least one energy storage element of at least one first module relative to the voltage of the at least one electrical energy storage element of at least one second module electrically connected to the at least one first module; determines a temporal proportion of magnetizing inductor path states and demagnetizing inductor path states; and selects corresponding inductor path states for the at least one first and the at least one second module based on the power path states of the at least one first module and the at least one second module with a lookup table.

Cited non-patent literature

• Wang, Boshuo, Zhongxi Li, Charles E. Sebesta, Daniel Torres Hinojosa, Qingbo Zhang, Jacob Robinson, Gang Bao, Angel V. Peterchev, and Stefan Goetz (2022). “Multichannel power electronics and magnetic nanoparticles for selective thermal magnetogenetics.” Journal of Neural Engineering, doi: 10.1088/1741-2552/ac5b94
• Li, Zhongxi, Ricardo Lizana, Zhujun Yu, Sha Sha, Angel V. Peterchev, and Stefan M. Goetz (2019). “A modular multilevel series/parallel converter for a wide frequency range operation.” IEEE Transactions on Power Electronics, 34(10): 9854-9865. doi: 10.1109/TPEL.2019.2891052
• Goetz, Stefan M., Zhongxi Li, Xinyu Liang, Chengduo Zhang, Srdjan M. Lukic, and Angel V. Peterchev (2016). “Control of modular multilevel converter with parallel connectivity-Application to battery systems.” IEEE Transactions on Power Electronics, 32(11):8381-8392, doi: 10.1109/TPEL.2016.2645884
• Li, Zhongxi, Jinshui Zhang, Angel Peterchev, and Stefan Goetz (2022). Modular Pulse Synthesizer for Transcranial Magnetic Stimulation with Flexible User-Defined Pulse Shaping and Rapidly Changing Pulses in Sequences. arXiv preprint arXiv:2202.06530
• Zeng, Zhiyong, Lari Koponen, Rena Hamdan, Zhongxi Li, Stefan Goetz, and Angel V. Peterchev (2022). “Modular multi-level TMS device with wide output ranges and ultrabrief pulse capability for sound reduction.” Journal of Neural Engineering, 19(2): 026008, doi: 10.1088/1741-2552/ac572c
• Goetz, Stefan M., Michael Pfaeffl, Jonas Huber, Matthias Singer, Rainer Marquardt, and Thomas Weyh (2012). “Circuit topology and control principle for a first magnetic stimulator with fully controllable waveform.” In 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 4700-4703. doi:' 10.1109/EMBC.2012.6347016
• Fang, Jingyang, Frede GE Blaabjerg, Steven Liu, and Stefan Goetz (2021). “A review of multilevel converters with parallel connectivity.” IEEE Transactions on Power Electronics, 36(11):12468-12489. doi: 10.1109/TPEL.2021.3075211

QUOTES INCLUDED IN THE DESCRIPTION

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

Cited patent literature

• US 9502960 [0004]
• US 9496799 [0004]
• US 5963086 [0007]
• US 8149061 [0008]
• US 8604883 [0009]
• US 8441224 B2 [0024]
• US 7279855 B2 [0024]
• US 8760122 B2 [0024]
• US 20190288617 A1 [0024]

Claims (10)

electrical circuit, comprising at least three modules, each the modules include at least one energy storage element and at least three electrical switches, each of the modules is electrically connected to at least a second of the modules by at least one intermodule electrical connection, and the at least two electrical switches in each module are configured to provide an electrical connection between a first module and at least a second module between at least two alternatives of electrical connection by effectively switching the electrical switches to change and wherein at least two modules are connected by at least two electrical connections, at least one of which has inductance of more than 100 nH. Electrical circuit according to Claim 1 , wherein the at least two alternatives of electrical interconnection enable at least one series state in which the at least one energy storage element of a first module is connected in series with at least one energy storage element of at least a second module, and the at least two alternatives of electrical connectivity provide at least one bypass Enable state in which at least one energy storage element of a first module is electrically connected to a second module with at most one of its electrical contacts. Electrical circuit according to Claim 2 , characterized in that at least a part of the paired electrical connection of at least one module of each branch is connected to at least one positive connection of the energy storage elements of the at least one module of each branch and at least another part of the paired electrical connection of at least one module of each branch is connected to at least one negative Connection of the energy storage elements of the at least one module of each branch is connected so that the energy storage elements are permanently connected in parallel or into a unit. Electrical circuit according to Claim 3 , wherein the at least three electrical switches of a first module, together with the at least three electrical switches of a second module that is electrically connected to the first module, can dynamically switch between a plurality of connection states determined by a combination of a plurality of power path states and a plurality of inductor path states. States are formed, and wherein the connection states include at least four power path states and at least four inductor path states, which can be combined by appropriately switching the at least three electrical switches on and off. Electrical circuit according to Claim 1 , in which a DC power supply is connected to the at least one energy storage element of at least one module, so that the DC power supply is electrically connected in parallel to the energy storage element. Method for generating at least one electrical output voltage of high power and quality by an electrical circuit Claim 1 , wherein the method performs at least the following steps: an output generating step that determines the power states of multiple modules such that the connection of the energy storage elements of the corresponding modules, each charged to an individual voltage level, corresponds to the selected power path states an approximation to the requested Output voltage generated with an error of at most the smallest voltage of one of the energy module elements of the modules; and a capacitor voltage charging step that receives the selected power path states of the modules to select an inductor path state for each of at least a first and at least a second module, the selected inductor path state magnetizing the at least one inductor or the at least one inductor that is the at least a first and the at least one second module connects, demagnetized. Procedure according to Claim 6 , in which the magnetization of the at least one inductor is achieved by combining at least one power path state selected from at least series+, series-, bypassL and bypassH with at least one inductor path state selected from I+, I-, IL, IH is achieved and at least temporarily maintains the magnetization of the at least one inductor by combining power path states with inductor path states that result in approximately constant magnetization and include at least one of the following combinations: the combination of Series+ and I+; the combination of series and I; the combination of BypassL and IL; the combination of BypassH and IH. Procedure according to Claim 7 , wherein the method changes the magnetization of the at least one inductor by selecting any power-inductor path combinations that produce a voltage across the at least one inductor by connecting at least one energy storage device in a circuit to the at least one inductor. Procedure according to Claim 8 , wherein the ratio of the durations of at least one magnetizing inductor path state and the demagnetizing inductor path state between at least a first and at least a second module is selected by an affine linear relationship with the desired voltage ratio between the energy storage elements of the at least first and at least second modules. Procedure according to Claim 9 , wherein the capacitor voltage charging step determines the inductor path states by receiving the power path states from the output generating step; calculates a voltage ratio for the required target voltage of the at least one energy storage element of at least one first module relative to the voltage of the at least one electrical energy storage element of at least one second module electrically connected to the at least one first module; determines a temporal proportion of magnetizing inductor path states and demagnetizing inductor path states; and selects corresponding inductor path states for the at least one first and the at least one second module based on the power path states of the at least one first module and the at least one second module with a lookup table.

Images