CONTROL METHOD OF A BOOST CONVERTER WITH N CONVERTER CELLS

Description

[0001] The present invention relates to a method for controlling a type converter Boost to N switching cells.

[0002] The invention relates to a Boost type DC / DC converter, known in English as Step-Up, which is a DC voltage boost converter.

[0003] In the field of motor vehicles, hybrid or electric, it is known to use Boost converters 2 in an electrical assembly 10 comprising an electric traction machine 4, as shown in figure 1 .

[0004] In such an electrical assembly 10 the converter Boost 2 is placed upstream of the voltage inverter 3 and the electric machine 4, considered as being a dynamic current source.

[0005] The converter Boost 2 is however positioned downstream of the electrical source 1 and the input filters.

[0006] A Boost 2 converter is a DC-DC converter. Its purpose is to provide the load it powers with a DC voltage greater than or equal to that measured at the input.

[0007] This type of converter is a non-minimum phase system. Indeed, the transfer function that connects the output voltage to the control, denoted H(s), presents in the complex plane, a zero with a positive real part.

[0008] According to a small-signal model development and linearization around a given stabilized operating point, the transfer function H(s) has the following form: H s = v ˜ out d ˜ = s − z 1 s − p 1 s − p 2

[0009] With z 1 >0, p 1 <0 and p 2 <0 are the zero and poles of the transfer function H(s) respectively.

[0010] On the other hand, a Boost 2 converter is a non-minimum phase system that has an output response that is initially inverse to the variation of the input. The output tends to diverge before converging towards the stabilized point. This is called a dropout phase of the output voltage. In addition, the relationships between, on the one hand, its different states and, on the other hand, the control is non-linear.

[0011] Also a general problem is to control converters stably and quickly Boost to N switching cells Cell 1 -Cell N , with N a non-zero natural integer, in particular a converter Boost synchronous type, as represented in figure 2 , in which the switching cells are composed of semiconductors, for example MOSFETS, as opposed to the passive Boost converter comprising diodes.

[0012] In particular, the prior art includes document US2017 / 0257038A1 which describes a method for controlling a converter. Boost N-phase, with N a non-zero natural integer. This method aims to modify the switching frequency of the system and adds a phase shift between the commands of the different cells in order to avoid the resonance frequency of the DC bus. However, such a solution does not allow for stable, dynamic and fast control. The following document is also known which discloses boost-type DC-DC converters: H. Renaudineau et al., "Efficiency Optimization Through Current-Sharing for Parallel DC-DC Boost Converters With Parameter Estimation," in IEEE Transactions on Power Electronics, vol. 29, no. 2, pp. 759-767, Feb. 2014, doi: 10.1109 / TPEL.2013.2256369.

[0013] Also, the invention aims to make the control of a converter more stable and faster. Boost,and more particularly a synchronous Boost converter.

[0014] For this purpose, a method of controlling a converter is proposed. Boost synchronous with N switching cells, in which N is a non-zero natural integer, said converter Boost receiving as input a direct electrical voltage from a voltage source and providing to a dynamic load at output an output voltage greater than or equal to the input voltage, the method comprising: A step of acquiring the measurement values of said input and output voltages; A step of acquiring the input currents measured in each switching cell; The method also comprising: A step of estimating the total energy error in said Boost converter, as a function of said input voltage, said output voltage and said measured input currents; A step of estimating the variation of a power value at the output of the Boost converter as a function of said total energy error estimate in which the value of said power of the dynamic load at the output of the Boost converter (2) is estimated by proportional-integral regulation on said total energy error estimate; A step of estimating a value of said power of the dynamic load at the output of the Boost converter as a function of said estimate of the variation of the power value of the dynamic load at the output; A step of calculating the current setpoint for each switching cell, as a function of said value of said power of the dynamic load at the output;and A step of controlling each switching cell as a function of said calculated current setpoint, said step of estimating the total energy error (φ) in said Boost converter (2) being furthermore a function of the difference between the input power with the power of the dynamic load at the output (P out ) of the Boost converter (2) and a function of an output vector (y) defining a linear representation of said Boost converter, the total energy error being calculated according to the following equation: ; φ = ∫ τ V in i in − P ^ out ︸ y ^ dτ − y in which Vin is the input voltage, iin the input current such that Vin*iin represents the total power absorbed by the converter, Pout is the power of the dynamic load connected to the Boost converter (2) at the output, y is the output vector defining a linear representation of said Boost converter and representing the energy stored in the system, said output vector (y) is calculated as a function of a discrete sum, for each switching cell, of the product between the inductance value of said switching cell (Cell1 -CellN ) and the square of the input current measured in said corresponding switching cell (Cell1 -CellN ), and as a function of the product of an output capacitance value of the Boost converter (2) with the square of said output voltage.

[0015] Thus, the method makes it possible to obtain a command for which the estimated power converges towards the total power at the input of the power converter.

[0016] By this method, the output voltage of the converter can be regulated Boost to a set voltage imposed by a user while ensuring relatively good system stability and optimizing the reactivity and dynamics of the control with regard to external disturbances.

[0017] Thus, an estimation value with relatively rapid and stable convergence can be obtained.

[0018] Advantageously and in a non-limiting manner, the value of said power of the dynamic load at the output (P out ) of the Boost converter is estimated by proportional-integral regulation on said estimate of total energy error. Thus, an estimate value with relatively rapid and stable convergence can be obtained.

[0019] Advantageously and in a non-limiting manner, said step of estimating the total energy error in said Boost converter is a function of the difference between the input power and the power of the dynamic load at the output of the Boost converter and a function of an output vector defining a linear representation of said Boost converter. Thus, a relatively reliable estimate of the total energy error can be obtained.

[0020] This allows the output vector forming the linearization of the Boost converter to be determined relatively quickly and reliably.

[0021] The invention also relates to an electrical assembly comprising a direct current electrical source, a Boost converter with N switching cells, a direct-alternating voltage converter, an electrical machine, and a device for controlling said Boost converter adapted to implement a method as described previously.

[0022] The invention also relates to a motor vehicle comprising such an electrical assembly.

[0023] Other features and advantages of the invention will emerge from reading the description given below of a particular embodiment of the invention, given for informational purposes but not as a limitation, with reference to the appended drawings in which: [ Fig. 1 ] is a principle view of a power supply assembly according to the invention; [ Fig. 2 ] is a schematic view of a converter Boost to N phases, N being a non-zero natural integer, here greater than 3, controlled according to the first embodiment of the invention [ Fig. 3 ] is a schematic view for a switching cell of the Boost converter and equivalent schematics for each of its operating modes; [ Fig. 4] is a synoptic view of the estimator of the input current setpoint of each switching cell of the Boost converter.

[0024] In a first embodiment of the invention, with reference to the figures 1 to 4 , we describe a non-linear control method by exact linearization with state feedback of a voltage boost converter to " N » Cell 1 -Cell N switching cells.

[0025] In reference to the figure 2 , each switching cell Cell 1 -Cell N comprises two controlled switches S 1 -SN and S 1- SN and an inductance L 1 -LN .

[0026] This process makes it possible to compensate for the non-linearities presented by the system that we wish to control.

[0027] The form of this compensation is linked to the structure of the system. For example, to compensate a non-linear term, denoted ϑ(x), by a subtraction, the control vector u at the input of the system and the non-linear term ϑ(x) must appear together in the form of an addition u+ϑ(x).

[0028] On the other hand, to compensate ϑ(x), by a division, the control vector u and the non-linear term must appear together in the form of a multiplication u·ϑ(x).

[0029] The converter equations Boost 2 to N switching cells are as follows: C out dV out dt = 1 − α 1 i 1 + 1 − α 2 i 2 + ⋯ + 1 − α N i N − P out V out L 1 di 1 dt = V in − 1 − α 1 V out L 2 di 2 dt = V in − 1 − α 2 V out ⋮ L N di N dt = V in − 1 − α N V out

[0030] Equation (1) can be written in the generic form of a nonlinear affine equation as follows: dX dt = f X t + g X t ⋅ u t

[0031] Thus, equation (2) becomes: di 1 dt di z dt ⋮ di N dt dV out dt = V in − V out L 1 V in − V out L 2 ⋮ V in − V out L N 1 C out ∑ k = 1 N i k − P out V out + V out L 1 0 ⋯ 0 0 V out L 2 ⋯ 0 0 0 ⋱ 0 0 0 ⋯ V out L N − i 1 C out − i 2 C out ⋯ − i N C out α 1 α 2 ⋮ α N with X∈R n< =(i 1 i 2 ... i NV out ) t< the vector of the state variables of the system, u(t)=(α 1 ··· α N )∈R m< represents the control vector considered as an input to the system, α k being the duty cycle applied to the switch S k of the kth< switching cell, and 1- α k the duty cycle applied to the switch S k of the kth switching cell. f(X) and g(X) are the indefinitely differentiable nonlinear functions; (L 1 ,L 2 ,...,LN ) and C out are the inductances of each switching cell and the capacitor at the output of the converter Boost respectively; i 1 , i 2 , ...,i N and V out are the currents in each inductance and the output voltage of the converter Boost 2 respectively; P out is the power of the dynamic load connected to the converter Boost 2 out.

[0032] In the following description, with reference to the Figures 3 And 4 , the converter Boost2 will only be described for a single switching cell.

[0033] The converter Boost 2, also called Boost 2, has two operating modes: a first mode 31 in which the inductance L 1 of the Boost 2 charges while the capacitor C out at the output discharges; and a second mode 32 in which the inductance L 1 of Boost 2 discharges while the capacitor C out at the output charges.

[0034] The first mode 31 is governed by the following electrical equations: di 1 dt = V in L 1 dV out dt = − 1 C out ⋅ P out V out

[0035] Concerning the energy variation in this first mode 31, the equation of the energy stored in the inductance is of the form E L = 1 2 L 1 i 1 2 . Thus, the temporal variation of the latter takes the following form: dE L dt = dE L di 1 ⋅ di 1 dt = V in i 1

[0036] Furthermore, the equation for the energy stored in the capacitor is of the form E C = 1 2 C out V out 2 . Thus, the temporal variation of the latter takes the following form: dE C dt = dE C dV out ⋅ dV out dt = − V out i out

[0037] The second mode 32 is governed by the following electrical equations: di 1 dt = 1 L 1 V in − V out dV out dt = 1 C out i 1 − P out V out

[0038] In terms of energy variation in this mode, we have: The time variation of the energy stored in the inductance takes the following form: dE L dt = dE L di 1 di 1 dt = V in − V out i 1 The time variation of the energy stored in the output capacitor takes the following form: dE C dt = dE C dV out dV out dt = i 1 − i out V out

[0039] Therefore, the variation of the total energy stored in the system, denoted E< t, is as follows: dE t dt = d E L + E C dt = V in i 1 − P out And : dE t dt = pertes dans le système

[0040] Generally speaking, the following output vector allows a global linearization of the system for the case of a converter with N switching cells: <menclose notation="box"> y = h X t = 1 2 ∑ k = 1 N L k i k 2 + 1 2 C out V out 2 < / menclose>

[0041] A first-order derivative for equation (12) gives the following equation: y ˙ = V in ∑ k = 1 N i k − P out

[0042] Now equation (13) corresponds to the variation of the total energy in the system as a function of time, which represents the losses dissipated by the converter.

[0043] We seek to estimate the total current setpoint i in ref absorbed by the DC / DC converter. However, for a power converter with efficiency η and according to the theorem of conservation of power we obtain: η V in i in ︸ P in = V out i out ︸ P out

[0044] With : η = P out P in ≤ 1 And i in = ∑ k = 1 N i k .

[0045] Thus, the input current becomes equal to: i in = P in V in = 1 η ⋅ P out V in

[0046] Consider the following system of equations: P ˙ out = σ P ¨ out = σ ˙ with P min ≤ P out = ∫ σ ⋅ dτ ≤ P max

[0047] We distinguish the following two cases: in transient mode: σ≠0 and P out is variable; in steady mode: σ=0 and P out is constant

[0048] On the other hand, let us take expression (13). The latter is composed of two terms: The first term V in ∑ k = 1 N i k = V in i in represents the total power absorbed by the converter. The input voltage and the current in the inductor are two variables physically measured using sensors. Thus, the total power absorbed by the converter is known.

[0049] The second term, P out , represents the useful power absorbed by the load. The latter is not measured but can be estimated. Thus, it is noted P̂ out .

[0050] From there, let the equation y ^ ˙ next: y ^ ˙ = V in i in − P ^ out

[0051] Thus, the estimation error will be: y ˙ − y ^ ˙ = − P out − P ^ out

[0052] When y ˙ − y ^ ˙ → 0 , we obtain: P ^ out → P out P ^ ˙ out = σ ^ → P ˙ out = σ P ^ ¨ = σ ^ ˙ → P ¨ out = 0

[0053] By expressing the equation P ˙ out = σ P ¨ out = σ ˙ (16) depending on the estimation error, we obtain: P ^ ˙ out = σ ^ − k 1 y ˙ − y ^ ˙ And P ^ ¨ out = − k 2 y ˙ − y ^ ˙ with k 1 and k 2 the observer's gains to be determined.

[0054] We note that these equations take the form of a Luenberger type observer where the estimation error term is included y ˙ − y ^ ˙ .

[0055] The purpose of this term is to drive the estimated state towards the actual state over time.

[0056] For equation (22), and after an adequate development, we obtain: <menclose notation="box"> P ^ ˙ out = k 2 ∫ τ V in i in − P ^ out ︸ y ^ dτ − y ︸ φ < / menclose>

[0057] For equation (21), and after an adequate development, we obtain: <menclose notation="box"> P ^ out = k 1 φ + k 2 ∫ τ φ dτ ︸ P ^ out < / menclose> with φ = ∫ τ V in i in − P ^ out ︸ y ^ dτ − y

[0058] Equation (25) represents the total energy error in the power converter, i.e. the deviation between the output y and its estimate, y being the vector representing the energy stored in the system.

[0059] Equation (2) allows us to determine that the estimated power is obtained using a Proportional-Integral type regulator which acts on the error φ.

[0060] In steady state, a stable asymptotic convergence of the error φ towards zero implies the following: ∫ τ V in i in − P ^ out dτ → en régime stabilisé y with y = 1 2 ∑ k = 1 N L k i k 2 + 1 2 C out V out 2 = cte ≠ 0 .

[0061] Also, for a given operating point: ∫ τ V in i in − P ^ out dτ → en régime stabilisé cte ≠ 0

[0062] The derivative of equation (27) implies: V in i in − P ^ out → en régime stabilisé 0

[0063] And therefore: <menclose notation="box"> P ^ out → en régime stabilisé V in i in = P in < / menclose>

[0064] By this approach, the estimated power converges to the total input power, which is absorbed by the power converter.

[0065] Thus, the notion of system efficiency disappears with this approach and the input reference current converges towards the following expression: i in ref → en régime stabilisé P in V in

[0066] Consequently, the setpoint current in each switching cell “k” is equal to: <menclose notation="box"> i k ref = i in ref N < / menclose> The observer's gains k 1 and k 2 are then determined from the equation k 1 = 2 ξw n k 2 = w n 2 (32) following: k 1 = 2 ξw n k 2 = w n 2

[0067] With wn =2πf n (the bandwidth of the estimator) with fn =[100Hz;600Hz] for example and ξ>0 (damping factor).

[0068] A stable asymptotic convergence of the error φ towards zero is then obtained for positive values of k 1 and k 2 .

Claims

1. Method for controlling a synchronous boost converter (2) with N switching cells (Cell1-CellN), in which N is a nonzero natural number, said boost converter (2) receiving at the input a DC electric voltage (Vin) from a voltage source and providing to a dynamic load, at the output, an output voltage (Vout) greater than or equal to the input voltage (Vin), the method comprising: - a step for acquiring values of the measurements of said input (Vin) and output (Vout) voltages; - a step for acquiring input currents measured in each switching cell; - a step for estimating the total energy error (φ) in said boost converter (2), as a function of said input voltage, said output voltage and said measured input currents; - a step for estimating the variation in a power value (Pout) of the dynamic load at the output of the boost converter as a function of said estimate of the total energy error (φ); - a step for estimating a value of said power (Pout) of the dynamic load at the output of the boost converter as a function of said estimate of the total energy error (φ) in which the value of said power of the dynamic load at the output of the boost converter (2) is estimated by proportional-integral control on said estimate of the total energy error; - a step for calculating the current setpoint for each switching cell (Cell1-CellN), as a function of said value of said power (Pout) of the dynamic load at the output; and - a step for controlling each switching cell (Cell1-CellN), as a function of said calculated current setpoint; said step for estimating the total energy error (φ) in said boost converter (2) being additionally a function of the difference between the input power and the power (Pout) of the dynamic load at the output of the boost converter (2) and a function of an output vector (y) defining a linear representation of said boost converter, the total energy error being calculated according to the following equation: φ = ∫ τ V in i in − P ^ out ︸ y ^ dτ − y in which Vin is the input voltage and iin is the input current such that Vin*iin represents the total power absorbed by the converter, Pout is the power of the dynamic load connected to the boost converter (2) at the output, y is the output vector defining a linear representation of said boost converter and representing the energy stored in the system, said output vector (y) is calculated as a function of a discrete sum, for each switching cell, of the product of the inductance value of said switching cell (Cell1-CellN) and the square of the input current measured in said corresponding switching cell (Cell1-CellN), and as a function of the product of an output capacitance value of the boost converter (2) and the square of said output voltage.

2. Method according to Claim 1, characterized in that the variation in a power value (Pout) of the dynamic load at the output of the boost converter is estimated by proportional control on said estimate of the total energy error.

3. Electrical assembly (10) comprising a DC electric source (1), a boost converter (2) with N switching cells, a DC / AC voltage converter (3), an electric machine (4), and a control device for said boost converter suitable for implementing a method according to either one of Claims 1 and 2.

4. Motor vehicle comprising an electrical assembly (10) according to Claim 3.