Capacitive-load driving circuit based on flying batteries

Abstract

The invention is a driving device with a network of N driving units, each having an energy-storing element and a switching matrix for series or parallel connection. A State of Charge mapping unit measures the state of charge for N elements, identifying a first set with low charge, and a second set. A control unit commands switching matrices through invest and recovery phases in cycles. In each cycle, a stacking sequence configures the network for capacitive load connection. The control unit calculates the sequence so that, during the recovery phase, elements in the first set receive charges from the capacitive load, previously supplied during the invest phase by elements in the second set.

Description

CAPACITIVE-LOAD DRIVING CIRCUIT BASED ON FLYING BATTERIES

[0001] Field of application

[0002] The present invention relates to the field devices (generally called converters) of powering active loads with reactive power or capacitive load based on batteries and more particularly batteries of rechargeable solid-state accumulators can be implemented into an integrated circuit on a semiconductor substrate. More specifically, the operation of powering according to the invention is based on an arrangement of phases during which one or more batteries are connected in series or parallel in order to form a cycle meeting a number of criteria.

[0003] Problem addressed

[0004] Soft robotics, micro-robotics, and haptics have shown promise in various applications, including computing interfaces and medicine. However, one of the significant challenges in these applications is the need for high driving voltages while also dealing with strict size and weight limitations.

[0005] Electrostatically-actuated transducers have the potential to benefit from higher voltages, as this can enhance their energy-density and force-displacement characteristics. Nevertheless, achieving these benefits at fixed electric field levels requires high driving voltages, which, in turn, pose significant challenges for power delivery circuits. In some cases, dielectric elastomer transducers, which are known for their relatively high force-displacement capabilities, may even require kilovolt-level driving voltages to achieve optimal performance. The common practice in high- voltage boost converters involves using magnetic components like inductors and transformers. However, these components do not scale well to smaller sizes, leading to reduced efficiency and power-density in miniaturized systems.

[0006] In the following description, it is meant by “reactive power” electrical power that oscillates back and forth between the source and the load without being consumed or dissipated as real power (also known as active power). Reactive power is associated with the reactive elements in an electrical circuit, primarily inductors and capacitors. More specifically, when a circuit contains capacitive loads, they store energy in an electric field when voltage is applied. This stored energy is alternately absorbed and released by the capacitive elements as the voltage across it varieswith time. This exchange of energy creates reactive power (meaning no power is dissipated in the reactive load), which is required to establish and modulate the electric field in the capacitive elements.

[0007] In the context of the invention, it is meant by “capacitive load” any load that behaves electrically like a capacitor. This is not limited to the capacitor component as such but can cover other types of load such as an electrostatic actuator. Capacitive loads, unlike conventional power electronic loads, require a significant amount of reactive power. This reactive power is needed to charge and discharge a large dielectric. In these systems, there is also a real (resistive) component accounting for losses and power delivered to the load. However, the real power is often overshadowed by the reactive power in this type of device. This makes traditional efficiency calculations challenging. Thus, an alternative metric Qx is used in order to evaluate drive system performance dominated by reactive power. This metric Qx calculates the ratio of total reactive power to real power loss in the circuit during a cycle (or period), assuming a periodic voltage signal is applied to the load. Qx can be viewed as an effective quality factor, indicating how well a circuit can supply and recover reactive energy while minimizing real power losses.

[0008] The invention intends to solve a technical problem in this field, related to the following aspects: increasing the robustness of the powering device delivering reactive power while increasing the quality factor Qx such as the real power loss tends to near zero values.

[0009] A more specific technical problem treated by the invention consists on is to extend the lifetime of the energy storage elements (herein named batteries) that form the power circuit. In the context of designing powering active loads with reactive power, it becomes necessary to monitor the state of charge of the batteries to ensure they discharge at a similar rate on average. This constraint arises because each battery serves as components in the driving circuit while also playing the role of an energy source (without any external input source).

[0010] Response to the problem and solution provided

[0011] The circuit according to the invention allows the battery to be used at the optimal operating point in terms of energy density. The charge balance of the entire closed cycle allows stable operation to be obtained, avoiding the need for additionalrecharging and discharging phases that would affect the overall efficiency of the driving circuit and its reliability.

[0012] Besides, the invention is based on adaptation of the stacking / unstacking sequence in order to control the state of charge of all the batteries of the driver during the exchanges with the load while maintaining the right output voltage pattern to drive the load accordingly. This allows the voltage variation across the terminals of the batteries to be minimized and therefore charge-sharing losses to be minimized. This allows the energy efficiency of the driver device to be increased.

[0013] Summary / Claims

[0014] The subject of the invention is a driving device for powering with reactive energy a load circuit comprising a capacitive load connected between an output node and a ground node. Said Driving device comprising:- a network of N driving units of rank i=1 to N, each unit comprising:• an energy-storing element , each consisting of at least one battery;• a switching matrix configured to connect said energy-storing element in series or in parallel between the output node and the ground node with the energy-storing element of an adjacent driving units ;- a State of Charge mapping unit configured to measure or estimate the state of charge of said N energy-storing elements in order to determine:• a first set of M1>1 energy-storing elements having a state of charge less than a first predetermined threshold;• a second set of energy-storing elements, of the chain, not belonging to the first set;- a control unit configured to command the switching matrices of said N driving units during a plurality of cycles ; each cycle comprising an invest phase followed by a recovery phase ; each phase consists of a stacking sequence corresponding to a series of connection configurations of the network of driving units to the capacitive load ; the control unit being configured furthermore to calculate said stacking sequence such as at the end of each cycle:• during the recovery phase , at least an energy-storing element belonging to the first set receives an amount of charges from the capacitive load ; said amount of charges being previously provided, during the invest phase , to the capacitive load by energy-storing elements belonging to the second set.

[0015] According to one particular aspect of the invention, during the invest phase: the number of connections of an energy-storing element belonging to the first set to output node is lower than the number of connections of an energy-storing element belonging to the second set to output node.

[0016] According to one particular aspect of the invention, during the recovery phase the number of connections of an energy-storing element belonging to the first set to output node is higher than the number of connections of an energy-storing element belonging to the second set to output node .

[0017] According to one particular aspect of the invention, the State of Charge mapping unit is configured to update periodically the composition of the first set and / or the second set by measuring the state of charge of said N energy-storing elements periodically after a predetermined number of cycles .

[0018] According to one particular aspect of the invention, the State of Charge mapping unit receives the calculated stacking sequence of a previous cycle via a feedback loop; said State of Charge mapping unit being configured to estimate, at the start of a later cycle, the evolution of state of charge of said N energy-storing elements starting from the combination of the measured state-of-charge of said previous cycle and said stacking sequence of said previous cycle.

[0019] According to one particular aspect of the invention, the second set comprises M2>1 of energy-storing elements having a state of charge higher than a second predetermined threshold.

[0020] According to one particular aspect of the invention, the stacking sequence keeps each energy-storing element operating between 10% and 90% of its state of charge.

[0021] According to one particular aspect of the invention, for each driving unit of rank i=1 a N-1 :- the energy-storing element has a first terminal and a second terminal;- the switching matrix comprises:• a first switch enabling to connect the first terminal of the energystoring element of rank i to the second terminal of the next energystoring element of rank i+1 ;• a second switch enabling to connect the first terminal of the energystoring element of rank i to the first terminal of the next energystoring element of rank i+1 ;• a third switch enabling to connect the second terminal of the energystoring element of rank i to the second terminal of the next energystoring element of rank i+1 .

[0022] According to one particular aspect of the invention, the driving device each driving unit comprises further an enabling switch linking the energy-storing element to the first switch and to the second switch .

[0023] According to one particular aspect of the invention, the batteries are solid-state batteries.

[0024] According to one particular aspect of the invention, the solid-state batteries are produced by stacking layers and comprise:- a positive electrode;- a negative electrode;- an inorganic solid electrolyte layer placed between the positive electrode and the negative electrode.

[0025] Brief description of the drawings

[0026] Other features and advantages of the present invention will become more clearly apparent upon reading the following description with reference to the following appended drawings.

[0027] [Fig. 1] Figure 1 illustrates a first curve P1 characteristic of the variation in the voltage across the terminals of a capacitor as a function of the amount of charge stored and a second curve P2 characteristic of the variation in the voltage across the terminals of a battery as a function of the amount of charge stored.

[0028] [Fig. 2] Figure 2 illustrates a circuit diagram of the network of driving units in the driving device according to the invention.

[0029] [Fig. 3] Figure 3 illustrates a circuit diagram of the driving device according to a first embodiment of the invention.

[0030] [Fig. 4] Figure 4 illustrates a circuit diagram of the driving device according to a second embodiment of the invention.

[0031] [Fig. 5a] Figure 5a illustrates of the invest phase of a first example of a powering cycle performed by the driving device according to the invention.

[0032] [Fig. 5b] Figure 5b illustrates of the recovery phase of a first example of a powering cycle performed by the driving device according to the invention.

[0033] [Fig. 6a] Figure 6a illustrates of the invest phase of a second example of a powering cycle performed by the driving device according to the invention.

[0034] [Fig. 6b] Figure 6b illustrates of the recovery phase of a second example of a powering cycle performed by the driving device according to the invention.

[0035] [Fig. 7] Figure 7 illustrates one example of a solid-state battery compatible with incorporation of the converter according to the invention into an integrated circuit.

[0036] The invention is specifically advantageous for battery-based implementations in comparison with capacitor-based implementations. To explain that, figure 1 illustrates a first curve P1 characteristic of the variation in the voltage across the terminals of a capacitor as a function of the amount of charge stored and a second curve P2 characteristic of the variation in the voltage across the terminals of a battery as a function of the amount of charge stored.

[0037] Both types of capacitive elements (capacitor and battery) are initially charged to an initial value Qinit thus setting the bias voltage of said element. In a capacitor, the voltage across its terminals is proportional to the charge stored on the positive terminal of the capacitor. Discharging the capacitor induces a decrease in the amount of charge, this causing a decrease in the voltage VCAP across the terminals of the capacitor. The slope of the variation in the voltage VCAP as a function of the charge stored QCAP is equal to the inverse of the capacitance of the capacitor. When the capacitor exchanges charge with the other elements of the converter, the value ofthe charge stored QCAP varies and the voltage VCAP varies in turn about the operating point with an amplitude of variation AVc.

[0038] In batteries, the relationship between the voltage across the terminals of the battery VBAT and the amount of charge stored QBAT in the battery is not a linear relationship. The curve P2 has three regions of operation:- a first region with a high increasing slope when the amount of charge QBAT is between 0 and a limit value QBI .- a second region in which the voltage plateau, in a range of variation in the charge bounded by two limit values QBI and QB2.- a third region with a high increasing slope when the amount of charge QBAT is between the limit value QB2 and the maximum charge value QBMAX.

[0039] From a practical point of view, there is no appreciable direct correlation between the voltage VBAT across the terminals of a battery and the amount of charge QBAT especially between QB1 and QB2, meaning that the obtained equivalent capacity tends to infinity. Specifically, in the plateau second region, the charge / discharge of the battery with respect to voltage has a very low slope meaning a small variation AVCin voltage for a wide range of variation in charge bounded by the two limit values QBI and QB2.

[0040] To obtain the same slope with a capacitor, a larger capacitor area would be required larger than that of a battery by 1000 to 10000 times. The low slope allows the amplitude of the variation AVc in the voltage across the terminals of the capacitive element to be limited. The decrease in AVJAQ increases the energy efficiency of the entire converter. Whence the advantage of using batteries in the driving device delivering reactive power.

[0041] The driving device according to the invention aims to exploit this low slope to improve, at equal area, the efficiency of said driving device by replacing the floating capacitors integrated into the converter with batteries, and by using a particular stacking / unstacking sequence. Positioning the operating point in the region of low slope further allows reducing the voltage rippleand thus induces less charge sharing loss at fixed frequency, and for a given amount of charge. Accordingly, the power efficiency of the circuit is improved. .

[0042] Thus, by comparing the curves P1 and P2, the advantage of using batteries instead of capacitors with regard to producing energy-storing elements in a driving device may be seen. Specifically, if batteries are used instead of capacitors, AVc will be greatly decreased for the same amount of charge exchanged (QCI -QC2) at equal area. In other words, at constant AVc between the two solutions, the amount of charge will be far higher (QCI -QC2 for the capacitor and QBI -QB2 for the battery). This improves the power efficiency of the circuit for a given at fixed frequency, and for a given amount of charge. .

[0043] In the context of the invention, the state of charge of each battery from the driving device must be kept in the range corresponding to the second region associated with the plateau in voltage as a function of charge.

[0044] Figure 2 illustrates a circuit diagram of the network of driving units Dili implemented in the driving device D1 according to the invention.

[0045] The driving device D1 according to the invention is intended for powering with reactive energy a capacitive load Cioad connected between an output node OUT and a ground node GND. The driving device D1 comprises a network of driving units DUi of rank i=1 to N. The capacitive load Cioad is connected to the driving unit DUN of rank i=N located at the end of the network. The driving device D1 is configured to drive with reactive power the capacitive load Cioad following a plurality of subsequent cycles. Each cycle consists of an invest phase followed by a recovery phase. Each phase comprises a stacking sequence corresponding to a series of connection configurations of the network of driving units DUi to the capacitive load Cioad. During the invest phase, the network of driving units DUi supply energy to the capacitive load Cioad. During the recovery phase, the capacitive load Cioad gives back the previously received energy to the network of driving units DUi supply energy.

[0046] The repetition of stacking sequence produces the application of a periodic signal on the output, generally speaking an alternative waveform at the resonant frequency of the electromechanical actuator comprising a capacitive load. The stacking sequence is set to provide this required voltage with some increasing (and decreasing) discrete steps, like N-level quantified signal. During the invest phase, the output signal Vout increases following a stepwise pattern. During the recovery phase the output signal Vout decreases following a stepwise pattern. The stepwise patterncan be periodic or non-periodic. The resulting rising time can be different from the resulting falling time. In the same way, the resulting rising slope can be different from the resulting falling slope. These parameters can be monitored by modifying the periodicity of the repetition of the stacking sequence and / or the duration of the recovery phase compared to the invest phase and vice versa.

[0047] In the context of the invention, the batteries serve two distinct purposes. Firstly, they provide energy to compensate for losses. Secondly, they incrementally increase the voltage, step by step to ensure charging. The more steps involved, the better the reduction in charge sharing losses. Additionally, the rechargeable nature of solid- state batteries allows them to recover the energy invested in the charging process. Consequently, the system is bidirectional and self-contained, meaning no external power source is required.

[0048] Each driving unit DUi comprises: an energy-storing element Bi and a switching matrix SMi. Each energy-storing element Bi consists of at least one battery. It is possible to produce energy-storing elements such that each consists of a single battery or of a plurality of batteries connected in series or in parallel depending on the optimal bias voltage and capacitance desired during the design of the energy-storing element. Advantageously, each energy-storing element Bi is one or more solid-state battery. A driving unit Dili can include further one or more capacitors mounted in parallel with the batteries forming together the energy-storing element of said driving unit DUi. Generally, the energy-storing elements are initially charged from an external source before starting the driving device D1. The energy-storing elements could be recharged after a given time of operation.

[0049] The switching matrix SMi is configured to connect said energy-storing element Bi in series or in parallel between the output node OUT and the ground node GND with the energy-storing element Bi+i, Bi-i of an adjacent driving units DUi+i or DUi-i. For each driving unit DUi of rank i=1 to N-1 , the switching matrix comprises a first switch Srev , a second switch Stop and a third switch Sbot . The first switch Srev enables to connect, directly or indirectly, the first terminal of the energy-storing element Bi of rank i to the second terminal of the next energy-storing element Bi+i of rank i+1. Thus, the activation of the first switch Srevj allows a connection in series between the associated energy-storing element Bi of rank i and the next energystoring element Bi+i of rank i+1. The second switch Stop enables to connect the firstterminal of the energy-storing element Bi of rank i to the first terminal of the next energy-storing element Bi+i of rank i+1 . The third switch Sbot enabling to connect the second terminal of the energy-storing element Bi of rank i to the second terminal of the next energy-storing element Bi+i of rank i+1. Thus, the simultaneous activation of the second switch Stop and the third switch Sbot allows a connection in series between the associated energy-storing element Bi of rank i and the next energystoring element Bi+i of rank i+1. This way, it is possible to determine the mounting configuration between the batteries of the network during each step of the sequence performed during an invest phase or a recovery phase. Each configuration corresponds to a predefined path followed by the charges from a predetermined group of batteries toward the capacitive load Cioad (during an invest phase) or inversely (during a recovery phase). As explained before, each cycle consists of an invest phase and a recovery phase. Each phase consists of a predetermined sequence of connection configurations of the network of driving units DUi regarding the capacitive load Cioad Accordingly, along multiple cycles, the energy is exchanged between the network of driving units Dili and the capacitive load Cioad without the need of an external supply voltage.

[0050] Note that the last driving unit Dlln ot rank i=N comprises the second and the third switches Stop and Sbotj and not the first switch Srevj as there is no next driving unit in this specific case.

[0051] Advantageously, each switching matrix SMi comprises further a central node Ni and an enabling switch Senj. The central node Ni is located between said enabling switch Senj and at least the second switch Stop of the associated switching matrix SMi. Besides, the switch Senj is mounted between the central node Ni and the first terminal of the associated energy-storing element Bi in a driving unit Dili. Thus, the enabling switch Senj allows selecting (connect or disconnect), independently of other driving units, the associated energy-storing element Bi in the network (or the stack).

[0052] According to one particular aspect of the invention, the first, second, third and enabling switches of each driving unit are produced with CMOS transistors.

[0053] The network of driving units Dili comprises further a first conduction line L1 connecting in series together all the second switches Stopj of the driving units forming the network. The extremity of the first conduction line L1 is connected to the outputnode OUT. Thus, the activation of at least a subset of the second switches Stop allows the creation of an electrical path towards the output node OUT. As a result, depending on the activation combination of the second switches Stop , charges can flow through at least a part of the first conduction line L1 from the selected batteries towards the output node OUT, to charge the capacitive load Cioad (during an invest phase) or inversely (during a recovery phase)

[0054] Moreover, The network of driving units DUi comprises further a second conduction line L2 connecting in series together all the third switches Sbot of the driving units forming the network. The extremity of the second conduction line L2 is connected to the ground node GND. Thus, the activation of at least a subset of the third switches Sbot allows the creation of an electrical path towards the ground node GND. The second connection line L2 enables to connect the output node OUT directly to the ground GND, or to connect ground to any negative electrode of a battery from the network.

[0055] Figure 3 illustrates a circuit diagram of the driving device D1 according to a first embodiment of the invention. The driving device D1 comprises a network of N driving units DUi ; a control unit SEQU configured to command the switching matrices of said N driving units DUi during a plurality of successive cycles Cyc ; and a State of Charge mapping unit SoCU configured to measure the state of charge of said N energy-storing elements Bi. By way of non-limitative illustration and without loss of generality, the network is composed of N=3 driving units.

[0056] The State of Charge mapping unit SoCU is configured to realize an initial measurement of the state of charge of each energy-storing elements Bi. This measurement allows the determination of:- a first set S1 of M1>1 energy-storing elements having a state of charge less than a first predetermined threshold;- and a second set S2 of energy-storing elements, of the network, not belonging to the first set.

[0057] For example, the initial measurement of the state of charge of all the energystoring elements Bi indicates the following elements: the state of charge of the first battery Bi is at 60%; the state of charge of the second battery B2 is at 80%, the state of charge of the third battery B3 is at 90%; the first predetermined threshold is set at70%. Thus, the first set S1 comprises the first battery Bi having a state of charge below the first threshold of 70%. The second set S2 is composed of the second battery B2 and the third battery B3.

[0058] Advantageously, the composition of the second set S2 is also determined with a second threshold of state of charge. In this case, the second set S2 comprises only the batteries having a state of charge higher than the second predetermined threshold. In the previous example, if the second threshold is set at 85% then the second set contains only the third battery B3.

[0059] Alternatively, the distribution of the batteries Bi on the first set S1 and / or the second set S2 can be on the basis of an increasing (or decreasing) ranking of the measured state of charges. For example, the first set S1 is composed of the battery presenting the lowest state of charge and all the other batteries are in the second set S2. Alternatively, the first set S1 is composed of the battery presenting the lowest state of charge and the second set S2 is composed of the battery presenting the highest state of charge.

[0060] In the State of Charge mapping unit, the condition of the battery (illustrated by their State of Charge) can be evaluated by looking at different parameters.

[0061] According to a first implementation, the State of Charge mapping unit is configured to measure the voltage at the terminals of an energy-storing element Bi If a certain relaxation time is ensured, there is a correlation between the measured voltage and the state of charge.

[0062] According to a second implementation, the State of Charge mapping unit is configured to measure the output impedance of an energy-storing element Bi. Various techniques allow the determination of impedance at a specific frequency or over a range. The value or shape of the measured output impedance provides information about the battery's state of charge. For example, the techniques cover methods like Discharging measuring, coulomb counting, open circuit voltage, impedance spectroscopy, impedance estimation, Kalman filter, AC internal resistance and fuzzy logic model.

[0063] According to a third implementation, the State of Charge mapping unit is configured to measure directly the state of charge using a potentiostat. A potentiostat is an instrument used in electrochemistry to control and measure the voltagedifference between a working electrode and a reference electrode. It allows for the study of the electrochemical behavior of an energy-storing element by applying a precise voltage and measuring the resulting current.

[0064] The different implementations described for the State of Charge mapping unit can be combined or used intermittently. This enables a more precise measure of the state of charge of the energy-storing elements Bi of the driving device.

[0065] The control unit SEQU receives the distribution of said first set S1 and second set S2 generated by the State of Charge mapping unit SoCU. Based on the composition of the first set S1 and second set S2, the control unit SEQU is configured to calculate the stacking / unstacking sequence used during the invest phase and the recovery phase of each cycle Cyc. The control unit SEQU applies control signals d1 ,d2,d3 to command the state of all the switches of the N driving units DUi forming the network. Each step of the sequence corresponds to a configuration of connection obtained by the control of the switches by control unit SEQU. Each phase (recovery or invest) consists of a specific sequence. Each cycle Cyc consists of an invest phase followed by a recovery phase. When the driving device D1 is powering the capacitive load Cioad, the cycle is repeated based on the calculated sequence and the reactive power is permanently exchanged between the energy-storing elements Bi and the capacitive load Cioad. The control unit SEQU can be implemented by a programmable sequencer.

[0066] The control unit SEQU is configured to calculate the stacking sequence such as at the end of each cycle: during the recovery phase, at least an energy-storing element Bi belonging to the first set S1 (group of batteries with low State of Charge) receives an amount of charges from the capacitive load Cioad. Said recovered amount of charges is previously provided, during the previous invest phase, to the capacitive load Cioad by energy-storing elements belonging to the second set S2 (group of batteries with high State of Charge).

[0067] For each powering cycle the sequence is calculated such as during the invest phase, the number of connections of every energy-storing element Bi belonging to the first set S1 to output node OUT is lower than the number of connections of every energy-storing element Bj belonging to the second set S2 to output node OUT.

[0068] For each powering cycle the sequence is calculated such as during the recovery phase, the number of connections of every energy-storing element Bi belonging to the first set S1 to output node OUT is higher than the number of connections of every energy-storing element Bj belonging to the second set S2 to output node OUT. Advantageously, for each powering cycle the sequence is calculated in order to maximize the number of used batteries belonging to the first set S1 during the invest phase (thus minimize the presence of S2 batteries), then, maximize the number of used batteries belonging to the second set S2 during the recovery phase.

[0069] The combination of State of charge mapping unit SoCU and the control unit SEQU according to the invention enables to generate a stacking sequence keeping each energy-storing element operating between 10% and 90% of its state of charge. In this case, all the batteries of the network are operating in the voltage plateau region. This enables to extend driving circuit lifetime compared with state-of-the-art solutions, and more specifically the lifetime of the batteries Bi. When a battery Bi is identified as presenting a low state of charge, this battery receives more charges than it gives during a cycle thanks to the calculated sequence based on state of charge mapping. This solution is particularly advantageous in the case of a battery network presenting a high variability rate. In that, due to variations in the manufacturing process, differential aging, or charge history, the batteries possess different characteristics such as storage capacity, internal resistance, and terminal voltage. The driving device D1 manages the distribution of effort from each battery to compensate for losses Eioss= PioSS / T, where T represents the cycle Cyc duration.

[0070] Therefore, the driving device D1 according to the invention allows manipulating the methods of stacking and unstacking batteries to distribute this effort between the batteries based on their initial state of charge or ongoing measurements throughout the driving device's D1 lifespan. By doing so, the system can effectively account for variations in battery characteristics and ensure a balanced effort distribution to compensate for losses.

[0071] According to a specific aspect of the invention, the state of charge mapping unit SoCU is configured to perform the measurements in an initial step before starting the reactive energy exchange between the driving device D1 and the capacitive load Cioad. Then, the control unit SEQU generates a stacking sequence on the basis of thedistribution of the batteries on the first set S1 and the second set S2. Then, the driving device D1 start powering the load during successive cycles following the calculated stacking sequence.

[0072] Alternatively, the state of charge mapping unit SoCll is configured to perform the measurements during intermittent state where the driving device’s operation is paused. Then, the control unit SEQll generates an updated stacking sequence on the basis of the new distribution of the batteries on the first set S1 and the second set S2.

[0073] Alternatively, the state of charge mapping unit SoCll is configured to perform measurements during operation for some or all the batteries. This allows for the detection of abnormal operation and the beginning of battery end-of-life (departure from the voltage plateau). These measurements can modify the previously defined sequence to reduce the demand on certain batteries or even disconnect them from the battery network during operation.

[0074] Alternatively, the state of charge mapping unit SoCll is configured to perform measurements periodically to update the composition of the first set and / or the second set S1 ,S2 every P cycles, where P is a predetermined number of cycles.

[0075] Figure 4 illustrates a circuit diagram of the driving device D1 according to a second embodiment of the invention. The technical features and advantages described for the first embodiment remain valid for the second embodiment. The second embodiment of the invention differs from the first embodiment in that it comprises further a feedback loop reinjecting the current stacking sequence into the state of charge mapping unit SoCll as input. In the second embodiment, the state of charge mapping unit SoCU is configured to estimate, at the start of a later cycle, the evolution of state of charge of said N energy-storing elements Bi starting from the combination of the measured state-of-charge of said previous cycle and said stacking sequence of said previous cycle. The second embodiment requires at least an initial measurement (during calibration or offline) performed by the state of charge mapping unit SoCU. The initial measurement is a starting point for the state of charge estimation. The State of Charge at the end of the sequence of a cycle is determined by adding the accumulated charge received or provided by each battery during the cycle. By integrating the entire series of cycles from the beginning of the poweringoperation, the current State of Charge distribution can be estimated. On the basis of the estimated State of Charge distribution, the sequence is updated in the same way described for the first embodiment.

[0076] However, it should be noted that this estimation becomes increasingly less accurate over time due to cumulative errors. To address this, a periodic State of Charge measurement can be employed to restart the estimation process and reset the cumulative errors. Additionally, the estimation can be enhanced by periodically sensing the voltage across each battery, as the charge given and received is influenced by battery voltages. Consequently, a combination of direct measurement and the estimation of the State of Charge using the historical sequence data can be employed to improve accuracy.

[0077] According to a further aspect of the invention compatible with the first and the second embodiments, the network of N driving units Dili may have more batteries Bi than the minimum required to generate the maximum output voltage. These additional batteries serve as backups. If the State of Charge mapping unit SoCU detects a very low state of charge for a certain battery Bn, the control unit SEQU can decide to remove said battery Bn from the stack and replace it with one of the backup batteries. Another scenario is when two batteries Bni and Bn2 are in poor condition. The control unit SEQU can choose to remove one battery Bni while keeping the other Bn2 in the stack to maximize the recharge of the first battery Bni. After a few cycles, the control unit SEQU can swap the battery roles by reintroducing the second battery Bn2 into the loop for recharging. Based on these examples, various strategies can be devised to maintain a constant average State of Charge while considering the design context, such as different battery capacities, different redox couples, and various types of batteries with different voltage plateaus.

[0078] For instance, if we consider different types of batteries with varying voltage plateaus, the threshold value for determining whether a battery is in a low (first set S1) or high (second state S2) state of charge is being adjusted accordingly.

[0079] According to a further aspect of the invention compatible with the first and the second embodiments, the control unit SEQU can also make decisions for the next sequence based on previous experience, similar to a machine learning process. Thealgorithm can be well-defined or derived from previous learning, where the weights are provided by the designer or adjusted during the driving device lifetime.

[0080] According to a further aspect of the invention compatible with the first and the second embodiments, the driving device D1 comprises further temperature sensors to adjust the voltage threshold of the energy-storing element Bi, as the voltage plateau is influenced by temperature.

[0081] Figure 5a illustrates of the invest phase Ph1 of a first example of a powering cycle Cyc performed by the driving device D1 according to the invention.

[0082] The described invest phase Ph1 consists of three consecutive steps (i), (ii) and (iii). Each step corresponds to a connection configuration of the batteries Bi , B2 and B3 obtained via the control of the switches of the associated switching matrices SM1, SM2 and SM3. The control of said switching matrices follows a sequence generated by the control unit SEQU on the basis of the State of Charge mapping.

[0083] In the following example, before starting the described cycle, State of Charge mapping unit SoCU measures QI>Q2>QS where Qi is the quantity of charge stored in the first battery Bi; Q2 is the quantity of charge stored in the second battery B2; Q3 is the quantity of charge stored in the third battery B3 The State of Charge mapping unit SoCU generates the following distribution: the first set S1 corresponding to the batteries showing a low state of charge comprises the third battery B3: S1 ={Bs}; the second set S2 corresponding to the batteries showing a high state of charge comprises the first and the second batteries Bi, B2 : S2={Bi B2}.

[0084] During the first step (i) of the invest phase Ph1 , the enabling switch Sen,i of the first battery Bi is in close state in order to connect said first battery to the powering network. The enabling switch Sen, 2 of the second battery B2 is in open state in order to disconnect said second battery to the powering network. The enabling switch Sen, 3 of the third battery B3 is in open state in order to disconnect said third battery to the powering network. The second switches Stop of the all driving units are in a close state. Thus, the first terminal of the first battery Bi is connected to the output node OUT through the formed first conduction line L1. This enables a parallel connection between the first battery Bi and the capacitive load Cioad connected between an output node OUT and a ground node GND. Thus, the first battery Bi delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad. The voltage on the output node OUTis equal to VB, where VB is the voltage of said first battery. To facilitate the comprehension of the example and without loss of generality, we consider in this example that the batteries voltages VB are equal.

[0085] During the second step (ii) of the invest phase Ph1 , the enabling switch Sen.i of the first battery Bi is in open state in order to disconnect said first battery to the powering network. The enabling switch Sen, 2 of the second battery B2 is in close state in order to connect said second battery to the powering network. The enabling switch Sen, 3 of the third battery B3 is in close state in order to connect said third battery to the powering network. The first switch Srev,2 of the second driving unit DU2 is in close state in order to connect in series the second battery B2 with the third battery B3. The third switch Sbot of the first driving unit DU1 is in close state in order to connect the second terminal of the second battery B2 to the ground node GND. The second switch Stop, 3 of the third driving unit DII3 is in close state in order to connect the first terminal of the third battery Bs to the output node OUT. Thus, the second battery B2 delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad and the third battery B3 delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad. The voltage on the output node OUT increases to 2.VB.

[0086] During the third step (iii) of the invest phase Phi , the enabling switch Sen.i of the first battery Bi is in close state in order to connect said first battery to the powering network. The enabling switch Sen, 2 of the second battery B2 is in close state in order to connect said second battery to the powering network. The enabling switch Sen, 3 of the third battery B3 is in close state in order to connect said third battery to the powering network. The first switch Srev.i of the first driving unit DU1 is in close state in order to connect in series the second battery B2 with the first battery Bi.The first switch Srev,2 of the second driving unit DU2 is in close state in order to connect in series the second battery B2 with the third battery B3. The second switch Stop, 3 of the third driving unit DU3 is in close state in order to connect the first terminal of the third battery B3 to the output node OUT. The equivalent circuit corresponds to the three batteries mounted in series between the output node OUT and the ground node GND. Thus, the first battery Bi delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad ; the second battery B2 delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad and the third battery B3 delivers a unit of charge +q=Cioad.VB to the capacitive load Cioad. The voltage on the output node OUT increases to 3.VB.

[0087] Figure 6b illustrates of the recovery phase Ph2 of the first example of a powering cycle Cyc performed by the driving device D1 according to the invention.

[0088] The described recovery phase Ph2 consists of three consecutive steps (i1), (ii’) and (iii’) performed by the driving device D1 following the previously described invest phase. Each step corresponds to a connection configuration of the batteries Bi, B2 and B3 obtained via the control of the switches of the associated switching matrices SM1, SM2 and SM3. The control of said switching matrices follows a sequence generated by the control unit SEQll on the basis of the State of Charge mapping.

[0089] During the first step (i1) of the recovery phase Ph2, the second and the third batteries B2 and B3 are both connected to the network via their respective enabling switches. The connection configuration is equivalent to a connection in series of the third and the second batteries B3 and B2, the whole mounted between the output node OUT and the ground node GND. Thus, the second battery B2 receives a unit of charge +q=Cioad.VB from the capacitive load Cioad and the third battery B3 receives a unit of charge +q=Cioad.VB from the capacitive load Cioad. The voltage on the output node OUT decreases to 2.\ / B.

[0090] During the second step (ii') of the recovery phase Ph2, only the third battery B3 is connected to the network via its enabling switches. The connection configuration is equivalent to a connection in parallel of the third battery B3 between the output node OUT and the ground node GND. Thus, the third battery B3 receives a unit of charge +q=Cioad.VB from the capacitive load Cioad. The voltage on the output node OUT decreases to VB.

[0091] During the third step (iii') of the recovery phase Ph2, all the batteries are disconnected from the network and the capacitive load Cioad is short-circuited to be discharged. The voltage on the output node OUT decreases to the ground voltage. Note thatafter the third step (iii1) of the recovery phase Ph2, the system has lost an energy quanta q due to the short-circuit of the capacitive load Cioad with the ground. Accordingly, the maximisation of the number of batteries in the network enables minimising the exchanged energy quanta q and thus, the lost energy quanta in the final step of each cycle. Besides, minimising the amplitude of the output voltage steps enables minimising the energy loss in the final step of each cycle.

[0092] Table 1 illustrates the cumulative variations in amounts of charge in the network complements during the execution of the cycle Cyc according to the first example.the batteries forming the network is reduced, as at the end of the cycle the State of Charge disparities in the network are reduced.

[0094] Figure 6a illustrates of the invest phase Ph1 of a second example of a powering cycle Cyc performed by the driving device D1 according to the invention.

[0095] The described invest phase Ph1 consists of three consecutive steps (i), (ii) and (iii). Each step corresponds to a connection configuration of the batteries Bi , B2 and B3 obtained via the control of the switches of the associated switching matrices SM1, SM2 and SM3. The control of said switching matrices follows a sequence generated by the control unit SEQll on the basis of the State of Charge mapping. The second example illustrates a sequence allowing a more precise tuning of the State of Charge during the invest phase.

[0096] In the following example, before starting the described cycle, State of Charge mapping unit SoCU measures QI>Q2>Q3. The State of Charge mapping unit SoCU generates the following distribution: the first set S1 corresponding to the batteries showing a low state of charge comprises the third battery B3: S1 ={Bs}; the second set S2 corresponding to the batteries showing a high state of charge comprises the first and the second batteries Bi , B2 : S2={Bi B2}.

[0097] During the first step (i) of the invest phase Ph1 , the first, second and the third batteries Bi, B2 and B3 are connected to the network via their respective enabling switches. The connection configuration is equivalent to a connection in parallel of the first , second and third batteries Bi, B2 and B3, the whole mounted between the output node OUT and the ground node GND. Thus, the first battery Bi delivers the amount of charge +q / 3=Cioad.VB / 3 to the capacitive load Cioad the second battery B2 delivers the amount of charge +q / 3=Cioad.VB / 3 to the capacitive load Cioad and the third battery B3 delivers the amount of charge +q / 3=Cioad.VB / 3 to the capacitive load Cioad. The total amount of charge received by capacitive load Cioad during said first step (i) is equal to +q=Cioad.VB. The voltage on the output node OUT increases from 0 to VB.

[0098] During the second step (ii) of the invest phase Ph1 , the first, second and the third batteries Bi, B2 and B3 are connected to the network via their respective enabling switches. The connection configuration is equivalent to a connection in parallel of the second and third batteries B2, B3. The combination of the two parallel batteries 62, B3 is mounted in series with the first battery Bi. The whole (B2 / / B3 in series with Bi) is mounted between the output node OUT and the ground node GND. Thus, the first battery Bi delivers the amount of charge +q=Cioad.VB to the capacitive load Cioad, the second battery B2 delivers the amount of charge +q / 2=Cioad.VB / 2 to the capacitive load Cioad and the third battery B3 delivers the amount of charge +q / 2=Cioad.VB / 2 to the capacitive load Cioad. The total amount of charge received by capacitive load Cioad during said second step (ii) is equal to +q=Cioad.VB. The voltage on the output node OUT increases from VB to 2.VB.

[0099] During the third step (iii) of the invest phase Ph1 , the first, second and the third batteries Bi, B2 and B3 are connected to the network via their respective enabling switches. The connection configuration is equivalent to a connection in series of the first, second and third batteries Bi , B2, B3. The whole is mounted between the output node OUT and the ground node GND. Thus, the first battery Bi delivers the amount of charge +q=Cioad.VB to the capacitive load Cioad, the second battery B2 delivers the amount of charge +q=Cioad.VB and the third battery B3 delivers the amount of charge +q=Cioad. B to the capacitive load Cioad. The total amount of charge received by capacitive load Cioad during said second step (ii) is equal to +q=Cioad.VB. The voltage on the output node OUT increases from 2.VB to 3.VB.

[0100] Figure 6b illustrates of the recovery phase Ph2 of the second example of a powering cycle Cyc performed by the driving device D1 according to the invention. The recovery phase of the second example is equivalent to the recovery phase of the first example.

[0101] Table 2 illustrates the cumulative variations in amounts of charge in the network complements during the execution of the cycle Cyc according to the second example.

[0102] As a result of the generated sequence, the gap between the states of charge of the batteries forming the network is reduced, as at the end of the cycle the State of Charge disparities in the network are reduced with a more precise tuning compared to the first example.

[0103] Figure 7 illustrates one example of a solid-state battery compatible with the implementation of the driving device according to the invention into an integrated circuit.

[0104] Solid-state batteries (or microbatteries) are a new battery technology in development. This technology is compatible with a microelectronic fabrication process. This type of microbattery has physical, density, thickness and unit-size characteristics allowing a compromise to be obtained between miniaturization andenergy efficiency. Said microbattery can be split into small battery without performance penalty.

[0105] Without loss of generality, a solid-state battery may consist of a stack of thin layers placed on a substrate SUB, which will generally be made of a semiconductor. The stack of thin layers comprises a positive electrode CAT made of a conductive material, a negative electrode AN made of a conductive material, and an inorganic solid-electrolyte layer ELY placed between the positive electrode and the negative electrode.

[0106] More specifically, a solid electrolyte capacitor consists of at least two electronically conductive electrodes, separated by a dielectric and ionically conductive material in the solid state, commonly known as the solid electrolyte. Several families of materials are considered for this role, including amorphous glasses (the best known being LiPON, which can be produced by several vacuum deposition techniques), crystallised oxides from the perovskite and garnet family (LLTO, LLZO, LATP), and sulphides (LGPS, LPS) for lithium solid electrolytes.These components are characterised by a dual capacitive response, the first in a low-frequency range due to the formation of an electrochemical double layer at the electrode / electrolyte interfaces, and the second due to the dielectric polarisation of the solid electrolyte. This specificity is characteristic of ionic capacitors using a solid electrolyte.

[0107] Such a structure is producible with a succession of technological manufacturing steps that are well-established in the semiconductor industry, such as chemical etching, plasma etching, sputtering and lithography.

[0108] With this type of technology, it is possible to produce batteries that have an area comprised between 100 pm2and 5 mm2, for an energy density ranging from 1 mAh / cm2to 10 mAh / cm2. By way of indication, at equal area, the equivalent of a battery having an energy density of 1 mAh / cm2is a capacitor having a capacitance per unit area of 9 mF / mm2. Currently, the best capacitances per unit area obtained with silicon-integratable capacitors are of the order of 1 pF / mm2. The use of solid- state batteries in the silicon-integrated converter thus allows the energy density of the storing elements to be multiplied by a factor of 9000 compared to the use of capacitors of equal area (and / or equal volume).

Claims

CLAIMS1. Driving device (D1) for powering with reactive energy a load circuit comprising a capacitive load (Cioad) connected between an output node (OUT) and a ground node (GND); said Driving device (D1) comprising:- a network of N driving units (DUi) of rank i=1 to N, each unit (DUi) comprising:• an energy-storing element (Bi), each consisting of at least one battery;• a switching matrix (SMi) configured to connect said energy-storing element (Bi) in series or in parallel between the output node (OUT) and the ground node (GND) with the energy-storing element (Bi+i , Bi-i) of an adjacent driving units (DUi+i DUi-i);- a State of Charge mapping unit (SoCU) configured to measure or estimate the state of charge of said N energy-storing elements (Bi) in order to determine:• a first set of M1>1 energy-storing elements having a state of charge less than a first predetermined threshold;• a second set of energy-storing elements, of the chain, not belonging to the first set;- a control unit (SEQU) configured to command the switching matrices of said N driving units (DUi) during a plurality of cycles (Cyc); each cycle comprising an invest phase (Ph1) followed by a recovery phase (Ph2); each phase (Ph1 , Ph2) consists of a stacking sequence corresponding to a series of connection configurations of the network of driving units (DUi) to the capacitive load (Cioad); the control unit being configured furthermore to calculate said stacking sequence such as at the end of each cycle:• during the recovery phase (Ph2), at least an energy-storing element belonging to the first set receives an amount of charges from the capacitive load (Cioad); said amount of charges being previously provided, during the invest phase (Ph1), to the capacitive load (Cioad) by energy-storing elements belonging to the second set.

2. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to claim 1 wherein during the invest phase (Ph1): the number of connections of an energy-storing element belonging to the first set to output node (OUT) is lower than the number of connections of an energystoring element belonging to the second set to output node (OUT).

3. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 or 2 wherein during the recovery phase (Ph2) the number of connections of an energy-storing element belonging to the first set to output node (OUT) is higher than the number of connections of an energystoring element belonging to the second set to output node (OUT).

4. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 3 wherein the State of Charge mapping unit (SoCU) is configured to update periodically the composition of the first set and / or the second set by measuring the state of charge of said N energy-storing elements periodically after a predetermined number of cycles (Cyc).

5. Driving device (D1 ) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 4 wherein the State of Charge mapping unit (SoCU) receives the calculated stacking sequence of a previous cycle via a feedback loop; said State of Charge mapping unit (SoCU) being configured to estimate, at the start of a later cycle, the evolution of state of charge of said N energy-storing elements (Bi) starting from the combination of the measured state- of-charge of said previous cycle and said stacking sequence of said previous cycle.

6. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 5 wherein the second set comprises M2>1 of energy-storing elements having a state of charge higher than a second predetermined threshold.

7. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 6 wherein the stacking sequence keeps each energy-storing element operating between 10% and 90% of its state of charge.

8. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 7 wherein for each driving unit (DUi) of rank i=1 a N-1 :- the energy-storing element (Bi) has a first terminal and a second terminal;- the switching matrix (SMi) comprises:• a first switch (Srev ) enabling to connect the first terminal of the energy-storing element (Bi) of rank i to the second terminal of the next energy-storing element (Bi+i) of rank i+1 ;• a second switch (Stopj) enabling to connect the first terminal of the energy-storing element (Bi) of rank i to the first terminal of the next energy-storing element (Bi+i ) of rank i+1 ;• a third switch (Sbot ) enabling to connect the second terminal of the energy-storing element (Bi) of rank i to the second terminal of the next energy-storing element (Bi+i ) of rank i+1 .

9. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to claim 8 wherein each driving unit (DUi) comprises further:• an enabling switch (Sen ) linking the energy-storing element (Bi) to the first switch (Srevj) and to the second switch (Stop ).

10. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to any one of claims 1 to 9 wherein the batteries are solid-state batteries.

11. Driving device (D1) for powering with reactive energy a capacitive load (Cioad) according to claim 10 wherein the solid-state batteries are produced by stacking layers and comprise:- a positive electrode;- a negative electrode;- an inorganic solid electrolyte layer placed between the positive electrode and the negative electrode.

Images