Multi-motor driving system torque optimal distribution method and system

Abstract

The invention relates to the technical field of electric driving system energy-saving control, and discloses a multi-motor driving system torque optimal distribution method and system, and the method comprises the steps: building a single power unit power loss model of a multi-motor driving system; constructing a fitting straight line segment; constructing a torque-current mapping relation curve; a train speed prediction model, a multi-motor driving system power unit motor rotating speed prediction model, a multi-motor driving system total torque reference value prediction model and a multi-motor driving system power unit power loss prediction model are constructed; establishing a multi-motor driving system energy efficiency evaluation function, a multi-motor driving system stability evaluation function, a multi-motor driving system total torque following evaluation function and a system optimization objective function; designing a learning rate of a control vector; and updating the control vector according to the learning rate, and obtaining an optimized torque distribution coefficient and an optimized torque reference value distributed by each power unit. Therefore, the torque optimization distribution of the multi-motor driving system is realized, and the energy utilization efficiency of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy-saving control of electric drive systems, and in particular to a method and system for optimizing torque distribution of a multi-motor drive system. Background Art

[0002] As the scale of high-speed railways continues to expand, the energy consumption problem in the railway industry has become increasingly prominent. Therefore, reducing energy consumption in the railway industry and promoting sustainable railway construction has become a top priority. High-speed trains belong to multi-motor axle-controlled drive systems, which are typical redundant drive systems in which each power unit can be controlled independently. Torque distribution is a unique technical advantage of redundant drive systems. It can set different torques for each power unit according to motor parameters and external working conditions while ensuring that the total output remains unchanged, so as to achieve better economic performance. However, although the existing fixed torque distribution strategy is simple to implement, it fails to fully tap the energy-saving potential of the train.

[0003] The current fixed allocation strategy has significant limitations in the following two aspects: First, during the operation of the train, due to the frequent low speed limit conditions, the power unit is in a low traction load working state, causing its working point to be far away from the high efficiency zone of the motor, thereby reducing the energy efficiency of the whole vehicle; second, the motor parameters of each power unit vary, resulting in differences in their efficiency characteristics and optimal working ranges. The fixed allocation strategy cannot ensure that all power units can work in the optimal position according to their efficiency characteristics. This allocation strategy that fails to fully consider the real-time changes in motor parameters is simple to implement, but it limits the energy-saving potential of the multi-motor drive system.

[0004] In order to further optimize the system energy efficiency, it is necessary to first consider the efficiency changes of the drive system under different speed and torque conditions, that is, it is required to establish a drive system loss estimation method with torque and speed as input. There are two main types of existing methods: one is to obtain the system efficiency distribution diagram based on finite element or experimental data. Although it can characterize the efficiency of the system in steady state, it is difficult to reflect its dynamic characteristics; the other type of method based on mechanism modeling is usually only applicable to surface-mounted permanent magnet synchronous motors, and is not applicable to more commonly used embedded motors. On the other hand, in the torque optimization allocation of multi-motor drive systems, the existing methods mainly include optimization methods based on fast search and heuristic algorithms. Although they are flexible, the calculation time increases significantly with the increase of control variables, which limits their practicality. Although the dynamic programming control algorithm can find the global optimal solution, the calculation amount is too large and lacks real-time performance. The quadratic programming method is suitable for specific optimization goals, but it has limited effect when dealing with complex optimization.

[0005] Based on the above problems, there is an urgent need for an online torque optimization allocation method for a multi-motor drive system that can fully consider the real-time changes of motor parameters and effectively improve the energy utilization efficiency of the system while ensuring system stability. Summary of the invention

[0006] The purpose of the present invention is to disclose a method and system for optimizing torque distribution of a multi-motor drive system, so as to solve the problem of difficulty in real-time estimation of system losses and online application of torque distribution in the process of energy efficiency optimization of a multi-motor drive system in the prior art, and is particularly suitable for improving the energy utilization efficiency of rail transit drive systems such as high-speed trains.

[0007] To achieve the above objectives, in a first aspect, the present invention discloses a method for optimizing torque distribution of a multi-motor drive system, comprising:

[0008] S1: establishing an inverter power device loss model and a permanent magnet synchronous motor power loss model respectively; and establishing a power loss model of a single power unit of a multi-motor drive system according to the inverter power device loss model and the permanent magnet synchronous motor power loss model;

[0009] S2: Use the torque and current sampling data of the power unit to build a data processing section;

[0010] S3: performing linear fitting on the data processing segment using the least square method to construct a fitting straight line segment; performing smoothing processing using a cubic B-spline interpolation method to construct a torque-current mapping relationship curve;

[0011] S4: construct prediction models for train speed, motor speed of a power unit of a multi-motor drive system, and total torque reference value of a multi-motor drive system respectively; construct a power loss prediction model for a power unit of a multi-motor drive system according to the torque-current mapping relationship curve and the power loss model of a single power unit of the multi-motor drive system;

[0012] S5: According to the power loss prediction model of the power unit of the multi-motor drive system, an energy efficiency evaluation function of the multi-motor drive system, a stability evaluation function of the multi-motor drive system and a total torque following evaluation function of the multi-motor drive system are respectively established; weight coefficients are respectively configured for the system energy efficiency evaluation function, the stability evaluation function and the total torque following evaluation function, and a system optimization objective function is established;

[0013] S6: Set the distribution coefficient of the torque reference value of each power unit as the control vector, calculate the gradient of the system optimization objective function; design the learning rate of the control vector according to the gradient of the system optimization objective function; set the initial solution of the control vector, update the control vector according to the learning rate, obtain the optimized torque distribution coefficient, and combine the total reference torque of the system to obtain the optimized torque reference value allocated to each power unit to achieve the optimal torque distribution of the multi-motor drive system.

[0014] In a second aspect, the present application also provides a control system for optimizing torque distribution of a multi-motor drive system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the processor implements the steps corresponding to the method described in the first aspect.

[0015] The present invention has the following beneficial effects:

[0016] The torque optimization distribution method for a multi-motor drive system provided by the present invention can not only comprehensively consider the impact of motor parameters and their changes on system performance, but also closely link the actual changes in motor parameters to accurately adjust the torque distribution coefficient in real time, effectively suppress the torque fluctuation problem occurring during the distribution process, ensure system stability and ride comfort, and at the same time reduce the total power loss of the multi-motor drive system and improve the energy utilization efficiency of the rail transit train drive system.

[0017] The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 It is a schematic diagram of the structure of a typical high-speed train multi-motor drive system applicable to the preferred embodiment of the present invention;

[0020] Figure 2 It is a flow chart of a method for optimizing control and allocating energy efficiency of a multi-motor drive system according to a preferred embodiment of the present invention;

[0021] Figure 3 It is a simulated traction operation cycle speed curve of a multi-motor drive system according to a preferred embodiment of the present invention;

[0022] Figure 4 It is a power unit torque reference value variation curve when the motor stator winding is increased by 30% in a preferred embodiment of the present invention;

[0023] Figure 5 is a total loss variation curve in a multi-motor drive system of a preferred embodiment of the present invention;

[0024] Figure 6 It is a percentage change curve of system loss reduction of the method proposed by the present invention relative to the torque average distribution method; DETAILED DESCRIPTION

[0025] The technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the usual meanings understood by persons with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not indicate quantity restrictions, but indicate the existence of at least one. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0027] It should be understood that the torque optimization distribution method of the multi-motor drive system provided in this application is applied to the rail transit traction control system, which includes 16 power units consisting of two-level three-phase inverters and permanent magnet synchronous motors in terms of hardware structure. Figure 1 As shown, this embodiment preferably adopts a multi-motor drive system for CRH3 high-speed trains, and the system adopts a multi-layer control structure; each power unit is controlled by its own torque control unit; the central control unit determines the total given torque according to the train running speed and the traction characteristic curve, and then distributes the set torque to 16 torque control units through the torque distribution link; finally, the motor uses the set torque as a control input to generate an electromagnetic torque to provide traction for the train.

[0028] like Figure 2 As shown, this embodiment provides a multi-motor drive system energy efficiency optimization control allocation method, including:

[0029] S1: establishing an inverter power device loss model and a permanent magnet synchronous motor power loss model respectively; and establishing a power loss model of a single power unit of a multi-motor drive system according to the inverter power device loss model and the permanent magnet synchronous motor power loss model;

[0030] S2: Use the torque and current sampling data of the power unit to build a data processing section;

[0031] S3: performing linear fitting on the data processing segment using the least square method to construct a fitting straight line segment; performing smoothing processing using a cubic B-spline interpolation method to construct a torque-current mapping relationship curve;

[0032] S4: construct prediction models for train speed, motor speed of a power unit of a multi-motor drive system, and total torque reference value of a multi-motor drive system respectively; construct a power loss prediction model for a power unit of a multi-motor drive system according to the torque-current mapping relationship curve and the power loss model of a single power unit of the multi-motor drive system;

[0033] S5: According to the power loss prediction model of the power unit of the multi-motor drive system, an energy efficiency evaluation function of the multi-motor drive system, a stability evaluation function of the multi-motor drive system and a total torque following evaluation function of the multi-motor drive system are respectively established; weight coefficients are respectively configured for the system energy efficiency evaluation function, the stability evaluation function and the total torque following evaluation function, and a system optimization objective function is established;

[0034] S6: Set the distribution coefficient of the torque reference value of each power unit as the control vector, calculate the gradient of the system optimization objective function; design the learning rate of the control vector according to the gradient of the system optimization objective function; set the initial solution of the control vector, update the control vector according to the learning rate, obtain the optimized torque distribution coefficient, and combine the total reference torque of the system to obtain the optimized torque reference value allocated to each power unit to achieve the optimal torque distribution of the multi-motor drive system.

[0035] It is worth noting that, in the present application, the inverter power devices include IGBTs and freewheeling diodes; the inverter power device loss model includes a conduction loss model and a switching loss model; wherein, the IGBT conduction loss and the freewheeling diode conduction loss include conduction voltage drop loss and conduction resistance loss; the IGBT switching loss and the freewheeling diode switching loss include turn-on loss and turn-off loss; the loss of the permanent magnet synchronous motor includes copper loss and iron loss; the power loss of a single power unit of the multi-motor drive system is composed of all IGBT conduction losses and switching losses of the inverter, all freewheeling diode conduction losses and switching losses, and the copper loss and iron loss of the permanent magnet synchronous motor.

[0036] This embodiment is based on the virtual simulation platform Simulink environment. The simulation platform consists of an inverter module, a permanent magnet synchronous motor module, a transmission mechanism module, a central control unit module, and a torque control unit module. The simulation platform is a commonly used existing technology in this field, and no further details are given here. The parameters used in the simulation experiment are shown in Table 1:

[0037] Table 1 Simulation experiment parameters

[0038]

[0039]

[0040] Optionally, the S1 includes:

[0041] S11: Establish a conduction loss model of a single IGBT and a single freewheeling diode of the inverter within one current cycle.

[0042] Calculate the duty cycle of the inverter IGBT at time t, the formula is:

[0043]

[0044] Where D Q (t) is the duty cycle of the IGBT at time t, V DC is the DC side voltage of the inverter, V s (t) is the inverter output phase voltage at time t, |V s (t)| is the inverter output phase voltage amplitude at time t, where V s (t) = V s cos(ω s t+φ), V s =|V s (t)| is the phase voltage amplitude, φ is the load power factor angle, ω s is the angular frequency;

[0045] The conduction loss model of a single IGBT of the inverter in one current cycle is established, and the formula is:

[0046]

[0047] Where P con_IGBT is the conduction loss of a single IGBT of the inverter in one current cycle; V t is the zero current voltage drop of IGBT, R ce is the equivalent resistance of the IGBT when it is turned on; I(t) is the current flowing through the IGBT at time t, I(t) = I s cos(w e ·t), where I s is the stator phase current amplitude, w e is the electrical angular velocity of the synchronous motor; T m is the current cycle; m is the modulation ratio of the inverter, m = V s / V DC ; The first part of the formula is the IGBT conduction voltage drop loss, where the second part is the on-resistance loss of IGBT;

[0048] Calculate the duty cycle of the inverter freewheeling diode using the formula:

[0049] D D (t) = 1-D Q (t) (3)

[0050] Where D D (t) is the duty cycle of the freewheeling diode connected in parallel with the IGBT in the inverter.

[0051] The conduction loss model of a single freewheeling diode of the inverter in one current cycle is established, and the formula is:

[0052]

[0053] Where P con_Diode is the conduction loss of a single freewheeling diode of the inverter in one current cycle, V f is the zero current voltage drop of the freewheeling diode, R ak is the equivalent resistance of the freewheeling diode when it is turned on; the first part of the formula is the conduction voltage drop loss of the freewheeling diode, where the second part is the on-resistance loss of the freewheeling diode.

[0054] S12: Establish a switching loss model of a single IGBT and a single freewheeling diode of the inverter within a current cycle.

[0055] The switching loss model of a single IGBT of the inverter in one current cycle is established, and the formula is:

[0056]

[0057] Where P sw_IGBT (t) is the switching loss of a single IGBT of the inverter in one current cycle; E on_IGBT (t) and E off_IGBT (t) are the turn-on energy loss and turn-off energy loss of the IGBT at time t; k on_IGBT and k off_IGBT They represent the mapping coefficients of the turn-on energy loss and turn-off energy loss of the IGBT and the current flowing through it; T m is a current cycle, where the first part is the IGBT turn-on loss, where the second part is the IGBT turn-off loss;

[0058] The switching loss model of a single freewheeling diode of the inverter in one current cycle is established, and the formula is:

[0059]

[0060] Where P sw_Diode (t) is the switching loss of a single freewheeling diode of the inverter in one current cycle; E on_Diode (t) and E off_Diode (t) are the turn-on energy loss and turn-off energy loss of the freewheeling diode at time t; kon_Diode and k off_Diode They represent the mapping coefficients of the turn-on energy loss and turn-off energy loss of the freewheeling diode and the current flowing through it; the first part of the formula is the turn-on loss of the freewheeling diode, where the second part is the turn-off loss of the freewheeling diode;

[0061] Formulas (2), (4), (5) and (6) constitute the inverter power device loss model (including the conduction loss model and switching loss model of the inverter IGBT and freewheeling diode).

[0062] S13: Establish a permanent magnet synchronous motor power loss model.

[0063] The voltage equation of the permanent magnet synchronous motor is established as follows:

[0064]

[0065] In the formula, R i is the equivalent iron loss resistance of the permanent magnet synchronous motor, R s is a single-phase stator winding, ψ f is the rotor permanent magnet flux, L d , L q are the d-axis and q-axis components of the stator winding inductance, u d 、u q are the d-axis and q-axis components of the motor stator voltage, respectively, i d 、i q are the d-axis and q-axis components of the motor stator current, i dt 、i qt are the torque current components of the motor stator current d and q axis components, i di 、i qi They are the iron loss current components of the d-axis and q-axis components of the motor stator current respectively.

[0066] According to the principle of power conservation, the input power of the permanent magnet synchronous motor is obtained as follows:

[0067] P in =u d i d +u q i q =P loss_Cu +P loss_Fe +P e (8)

[0068] Where P in is the input power of the permanent magnet synchronous motor, P loss_Cu and P loss_Fe are the copper loss and iron loss generated by the motor, P eis the electromagnetic power output by the motor, and the formulas are:

[0069]

[0070] P e =ω e ·T e =ω e ·(ψ f i qt +(L d -L q )i dt i qt )(11)

[0071] Where, T e is the output electromagnetic torque, ω e Electrical angular velocity of the traction motor.

[0072] The power loss model of permanent magnet synchronous motor is established, and the formula is:

[0073] P loss_PMSM =P loss_Cu +P loss_Fe (12)

[0074] Where P loss_PMSM is the power loss of the permanent magnet synchronous motor.

[0075] S14: Based on the above steps, a power loss model of a single power unit of a multi-motor drive system is established, and the formula is:

[0076] P loss_unit =6(P con_IGBT +P con_Diode +P sw_IGBT +P sw_Diode )+P loss_PMSM (13)

[0077] Where P loss_unit It is the power loss of a single power unit in a multi-motor drive system.

[0078] Optionally, S2 includes:

[0079] At fixed time intervals τ s The power unit torque and the corresponding d-axis and q-axis currents under the torque are continuously sampled, and each J groups of sampled data are constructed as a data processing segment, which is expressed as:

[0080] Q p ={S 1,p ,S 2,p ,...,S j,p ,...,S J,p}(14)

[0081] In the formula, Q p is the pth data processing segment, S j,p represents the jth group of data in the pth data processing segment, expressed as:

[0082] S j,p =(T j,p ,y j,p )(15)

[0083] Where, T j,p is the pth data processing segment Q p The power unit torque sampling data in the jth group of data, y j,p are the sampling data of the d-axis and q-axis components of the motor stator current in the j-th group of data in the p-th data processing segment, j={1,2,...,J}, p={1,2,...,P}, J is the total number of groups of sampling data in each data processing segment, and P is the total number of data processing segments.

[0084] Optionally, S3 includes:

[0085] S31: Use the least square method to perform linear fitting on the pth data processing segment, and the fitting coefficient satisfies the following relationship:

[0086]

[0087] In the formula, They are the first and second fitting coefficients of the pth data processing segment respectively.

[0088] Using the pth data processing segment Q p Construct a fitting straight line segment, the expression is:

[0089]

[0090] In the formula, is the pth data processing segment Q p The constructed fitting straight line segment, T p To fit a straight line segment The horizontal axis variable (power unit torque), To fit a straight line segment The ordinate variables (the d-axis and q-axis components of the motor stator current), that is, in the pth data processing section Q p Fitted straight line segment The horizontal axis value is T p (Power unit torque) corresponds to the ordinate value (d-axis and q-axis components of the motor stator current); and are the pth data processing segment Q p Minimum and maximum power unit torque sampling data.

[0091] S32: respectively process all P data processing segments {Q 1 ,Q 2 ,…,Q p ,…,Q P}Construct P fitting straight line segments The P fitting straight line segments are spliced; the spliced ​​P fitting straight line segments are smoothed using the cubic B-spline interpolation method to construct a torque-current mapping relationship curve, which is expressed as:

[0092]

[0093] In the formula, represents the torque-current mapping relationship curve constructed by smoothing P fitting straight line segments, T is the torque-current mapping relationship curve The horizontal axis variable (power unit torque), y is the torque-current mapping curve The ordinate variables (the d-axis and q-axis components of the motor stator current), that is, when y is y=i d and y=i q When , formula (18) is written as and

[0094] Optionally, S4 includes:

[0095] S41: Let the current time be t, consider the time interval between the sampling time and the control variable output, predict the train speed with a step length of τ, and build a train speed prediction model. The formula is:

[0096]

[0097] In the formula, v t+τ is the predicted train speed at time t+τ, v t is the train speed at time t, N is the total number of power units in the multi-motor drive system, mg is the positive pressure of the wheelset, τ is the prediction step, M is the mass of the train, r is the wheel diameter of the train, and f(·) represents the train resistance, which is a function of the train speed and mass. is the motor speed of the unit power unit at time t, is the adhesion coefficient of the unit power unit at time t, and the formula is:

[0098]

[0099] In the formula, c 1 、c 2 、c 3 、c 4They are the 1st, 2nd, 3rd and 4th parameters of the power unit adhesion coefficient, unit = 1, 2, ..., N.

[0100] S42: predicting the motor speed of the unit power unit of the multi-motor drive system with a step length of τ, and constructing a motor speed prediction model of the unit power unit of the multi-motor drive system, the formula is:

[0101]

[0102] In the formula, is the predicted motor speed of the unit power unit at time t+τ, η gear 、a gear Represent transmission efficiency and gear ratio respectively, J unit is the moment of inertia of the unit power unit, is the torque of the unit-th power unit at time t.

[0103] S43: predicting the total torque reference value of the multi-motor drive system with a step length of τ, and constructing a prediction model for the total torque reference value of the multi-motor drive system, the formula is:

[0104]

[0105] In the formula, is the predicted total torque reference value of the multi-motor drive system at time t+τ, Γ(·) represents the predicted train speed v at time t+τ t+τ The mapping relationship between the predicted total torque reference value of the multi-motor drive system at time t+τ can be obtained from the traction characteristic curve of the train.

[0106] S44: The power loss of the unit power unit of the multi-motor drive system is predicted with a step size of τ, and the current d-axis and q-axis components of the unit power at different torques are estimated using the established torque-current mapping curve. Combined with the power loss model of the drive system power unit established in S1, a power loss prediction model of the unit power unit of the multi-motor drive system is constructed. The formula is:

[0107]

[0108] In the formula, is the predicted power loss of the unit-th power unit of the multi-motor drive system at time t+τ, f I,P (·,·,·,) represents the mapping relationship between the current and power loss of the unit-th power unit in the multi-motor drive system. and In formula (18), y is y=i d and y=i qThe torque-current mapping curve established when It is the torque reference value assigned to the unit power unit at time t+τ, and the formula is:

[0109]

[0110] In the formula, is the distribution coefficient of the torque reference value of the unit power unit at time t+τ; unit=1,2,...,N; is the predicted total torque reference value of the multi-motor drive system at time t+τ.

[0111] Optionally, S5 includes:

[0112] S51: According to the power loss prediction model of the unit power unit of the multi-motor drive system, an energy efficiency evaluation function of the multi-motor drive system is established, and the formula is:

[0113]

[0114] In the formula, Represents the energy consumption evaluation function of the multi-motor drive system at the current moment.

[0115] S52: According to the motor speed prediction model of the unit power unit of the multi-motor drive system and the wheel slip rate, a stability evaluation function of the multi-motor drive system is established, and the formula is:

[0116]

[0117] In the formula, represents the stability evaluation function of the multi-motor drive system at the current moment, It is expressed as the slip rate of the unit power unit at time t+τ, and the formula is:

[0118]

[0119] It is the penalty coefficient of the slip rate of the unit power unit at time t+τ, and the formula is:

[0120]

[0121] In the formula, κ lim It is the limit value of the power unit slip ratio.

[0122] S53: According to the distribution coefficient of the torque reference value of the unit power unit, a total torque following evaluation function of the multi-motor drive system is established, and the formula is:

[0123]

[0124] In the formula, It is the total torque following evaluation function of multi-motor drive system.

[0125] S54: configuring weight coefficients for the three evaluation functions, and establishing a system optimization objective function:

[0126]

[0127] In the formula, ρ 1 , 2 and ρ 3 The evaluation functions are The weight coefficient of .

[0128] Optionally, S6 includes:

[0129] S61: Set the distribution coefficient of the torque reference value of each power unit at time t+τ to the control vector at the current time t The system optimization objective function is written as the control vector U t The functional form is denoted by J t =g(U t ), the system optimization objective function is in U t The gradient at is expressed as:

[0130]

[0131] In the formula, Indicates that the system optimization objective function is in U t The gradient at is the control vector U t The unit-th component in also represents the distribution coefficient of the unit-th power unit torque reference value at time t+τ; is the gradient operator; Indicates the control component of unit The gradient of is:

[0132]

[0133] In the formula, and Represent the current moment evaluation function and For the unit component at time t+τ number.

[0134] S62: Design the learning rate of the control vector according to the system optimization objective function gradient.

[0135] Design the learning rate of the unit-th control component, the formula is:

[0136]

[0137] In the formula, is the unit-th control component at the current time t The learning rate, A lim is the gradient limit of each iteration of the control component, α lim is the maximum learning rate of the control component set, when the gradient of the optimization objective function satisfies When the unit control component is With the set step size Perform gradient descent. When the gradient of the optimization objective function satisfies When the unit-th control component is With the set learning rate α lim Perform gradient descent.

[0138] unit=1,2,...,N, construct the learning rate vector of the system control vector at the current time t as

[0139] S63: Set the initial solution of the control vector to:

[0140]

[0141] S64: According to the learning rate vector, the unit-th control component of the control vector To update, the formula is:

[0142]

[0143] In the formula, and are the nth s +1st and nth s The unit-th control component in the control vector after iterations.

[0144] N s,unit Iterate until the learning rate of the unit-th control component meets the search stop condition: α end is the learning rate that satisfies the stop search condition; the control component obtained at this time That is, the current time t is N s,unit The unit-th optimized control component after the iterative update That is, the optimal distribution coefficient of the torque reference value of the unit power unit at time t+τ; combined with the total reference torque of the system According to formula (24), the optimized torque reference value allocated to the unit power unit is obtained as:

[0145] unit=1,2,...,N, the optimal distribution coefficient of the torque reference values ​​of all N power units at the current time t is obtained as follows: The optimized torque reference value distributed by all N power units is

[0146] Specifically, in one example, a high-speed train multi-motor drive system of this embodiment simulates a traction operation cycle speed curve as shown in FIG. Figure 3 The train starts to accelerate to a given speed of 140 km / h at 0 s, and enters a uniform speed condition at 1.7 s. The N (N=16) power units in the train are evenly divided into two groups, and the motor stator winding of one power unit is increased by 30%. Comparing the method proposed in the present invention with the currently used average distribution method, the torque reference value change curve of the power unit with a motor stator winding increased by 30% in the multi-motor drive system of this embodiment is shown in FIG. Figure 4 As shown; the total loss change curve of the multi-motor drive system is as follows Figure 5 As shown; the percentage change curve of the system loss reduction of the method proposed by the present invention relative to the torque average distribution method is as follows Figure 6 shown.

[0147] Corresponding to the above method embodiment, this embodiment provides a system for optimizing torque distribution of multi-motor drive control, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above method when executing the computer program. The system for optimizing torque distribution of multi-motor drive control can implement various embodiments of the above torque optimization distribution method for multi-motor drive system, and can achieve the same beneficial effects, which will not be described in detail here.

[0148] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for optimizing torque distribution of a multi-motor drive system, characterized in that: include: S1: establishing an inverter power device loss model and a permanent magnet synchronous motor power loss model respectively; and establishing a power loss model of a single power unit of a multi-motor drive system according to the inverter power device loss model and the permanent magnet synchronous motor power loss model; S2: Use the torque and current sampling data of the power unit to build a data processing section; S3: performing linear fitting on the data processing segment using the least square method to construct a fitting straight line segment; performing smoothing processing using a cubic B-spline interpolation method to construct a torque-current mapping relationship curve; S4: construct prediction models for train speed, motor speed of a power unit of a multi-motor drive system, and total torque reference value of a multi-motor drive system respectively; construct a power loss prediction model for a power unit of a multi-motor drive system according to the torque-current mapping relationship curve and the power loss model of a single power unit of the multi-motor drive system; S5: According to the power loss prediction model of the power unit of the multi-motor drive system, an energy efficiency evaluation function of the multi-motor drive system, a stability evaluation function of the multi-motor drive system and a total torque following evaluation function of the multi-motor drive system are respectively established; weight coefficients are respectively configured for the system energy efficiency evaluation function, the stability evaluation function and the total torque following evaluation function, and a system optimization objective function is established; S6: Set the distribution coefficient of the torque reference value of each power unit as the control vector, calculate the gradient of the system optimization objective function; design the learning rate of the control vector according to the gradient of the system optimization objective function; set the initial solution of the control vector, update the control vector according to the learning rate, obtain the optimized torque distribution coefficient, and combine the total reference torque of the system to obtain the optimized torque reference value allocated to each power unit to achieve the optimal torque distribution of the multi-motor drive system.

2. The torque optimization control distribution method for a multi-motor drive system according to claim 1, characterized in that: The inverter power device includes an IGBT and a freewheeling diode; the inverter power device loss model includes a conduction loss model and a switching loss model; wherein the IGBT conduction loss and the freewheeling diode conduction loss include conduction voltage drop loss and conduction resistance loss; the IGBT switching loss and the freewheeling diode switching loss include turn-on loss and turn-off loss; The losses of the permanent magnet synchronous motor include copper loss and iron loss; the power loss of a single power unit of the multi-motor drive system is composed of the conduction loss and switching loss of all IGBTs of the inverter, the conduction loss and switching loss of all freewheeling diodes, and the copper loss and iron loss of the permanent magnet synchronous motor.

3. The method for optimizing torque control and allocating for a multi-motor drive system according to any one of claims 1 or 2, characterized in that: The S1 includes: S11: Establish the conduction loss model of a single IGBT and a single freewheeling diode of the inverter in one current cycle as follows: Calculate the duty cycle of the inverter IGBT at time t to satisfy the following relationship: Where D Q (t) is the duty cycle of the IGBT at time t, V DC is the DC side voltage of the inverter, V s (t) is the inverter output phase voltage at time t, |V s (t)| is the inverter output phase voltage amplitude at time t, where V s (t) = V s cos(ω s t+φ), φ is the load power factor angle, ω s is the angular frequency; The conduction loss model of a single IGBT of the inverter in one current cycle is established to satisfy the following relationship: Where P con_IGBT is the conduction loss of a single IGBT of the inverter in one current cycle; V t is the zero current voltage drop of IGBT, R ce is the equivalent resistance of the IGBT when it is turned on; I(t) is the current flowing through the IGBT at time t, I(t) = I s cos(w e ·t), where I s is the stator phase current amplitude, w e is the electrical angular velocity of the synchronous motor; T m is the current cycle; m is the modulation ratio of the inverter, m = V s / V DC ; Calculate the duty cycle of the inverter freewheeling diode to satisfy the following relationship: D D (t)=1-D Q (t) (3) Where D D (t) is the duty cycle of the freewheeling diode connected in parallel with the IGBT in the inverter; The conduction loss model of a single freewheeling diode of the inverter in one current cycle is established to satisfy the following relationship: Where P con_Diode is the conduction loss of a single freewheeling diode of the inverter in one current cycle, V f is the zero current voltage drop of the freewheeling diode, R ak is the equivalent resistance of the freewheeling diode when it is turned on; S12: Establish the switching loss model of a single IGBT and a single freewheeling diode of the inverter in one current cycle as follows: The switching loss model of a single IGBT of the inverter in one current cycle is established to satisfy the following relationship: Where P sw_IGBT (t) is the switching loss of a single IGBT of the inverter in one current cycle; E on_IGBT (t) and E off_IGBT (t) are the turn-on energy loss and turn-off energy loss of the IGBT at time t; k on_IGBT and k off_IGBT They represent the mapping coefficients of the turn-on energy loss and turn-off energy loss of the IGBT and the current flowing through it, respectively. m is a current cycle; The switching loss model of a single freewheeling diode of the inverter in one current cycle is established to satisfy the following relationship: Where P sw_Diode (t) is the switching loss of a single freewheeling diode of the inverter in one current cycle; E on_Diode (t) and E off_Diode (t) are the turn-on energy loss and turn-off energy loss of the freewheeling diode at time t; k on_Diode and k off_Diode They respectively represent the mapping coefficients of the turn-on energy loss and turn-off energy loss of the freewheeling diode and the current flowing through it; The inverter power device loss model is constructed by formulas (2), (4), (5), and (6); S13: Establish a permanent magnet synchronous motor power loss model as follows: The voltage equation of the permanent magnet synchronous motor is established to satisfy the following relationship: In the formula, R i is the equivalent iron loss resistance of the permanent magnet synchronous motor, R s is a single-phase stator winding, ψ f is the rotor permanent magnet flux, L d , L q are the d-axis and q-axis components of the stator winding inductance, u d 、u q are the d-axis and q-axis components of the motor stator voltage, respectively, i d 、i q are the d-axis and q-axis components of the motor stator current, i dt 、i qt are the torque current components of the motor stator current d and q axis components, i di 、i qi are the iron loss current components of the d-axis and q-axis components of the motor stator current respectively; According to the principle of power conservation, the input power of the permanent magnet synchronous motor is obtained, which satisfies the following relationship: P in =u d i d +u q i q =P loss_Cu +P loss_Fe +P e (8) Where P in is the input power of the permanent magnet synchronous motor, P loss_Cu and P loss_Fe are the copper loss and iron loss generated by the motor, P e is the electromagnetic power output by the motor, satisfying the following relationship: P e =ω e ·T e =ω e ·(ψ f I qt +(L d -L q )i dt I qt ) (11) Where, T e is the output electromagnetic torque, ω e The electrical angular velocity of the traction motor; The power loss model of permanent magnet synchronous motor is established to satisfy the following relationship: P loss_PMSM =P loss_Cu +P loss_Fe (12) Where P loss_PMSM is the power loss of the permanent magnet synchronous motor; S14: Establish a power loss model for a single power unit of a multi-motor drive system to satisfy the following relationship: P loss_unit =6(P con_IGBT +P con_Diode +P sw_IGBT +P sw_Diode )+P loss_PMSM (13) Where P loss_unit It is the power loss of a single power unit in a multi-motor drive system.

4. The method for optimizing torque distribution of a multi-motor drive system according to claim 1, characterized in that: The S2 includes: At fixed time intervals τ s The power unit torque and the corresponding d-axis and q-axis currents under the torque are continuously sampled, and each J groups of sampled data are constructed as a data processing segment, satisfying the following relationship: Q p ={S 1,p ,S 2,p ,...,S j,p ,...,S J,p } (14) In the formula, Q p is the pth data processing segment, S j,p represents the jth group of data in the pth data processing segment, satisfying the following relationship: S j,p =(T j,p ,y j,p ) (15) Where, T j,p is the pth data processing segment Q p The power unit torque sampling data in the jth group of data, y j,p are the sampling data of the d-axis and q-axis components of the motor stator current in the j-th group of data in the p-th data processing segment, j={1,2,...,J}, p={1,2,...,P}, J is the total number of groups of sampling data in each data processing segment, and P is the total number of data processing segments.

5. The method for optimizing torque distribution of a multi-motor drive system according to claim 4, characterized in that: The S3 includes: S31: Use the least square method to perform linear fitting on the pth data processing segment, and the fitting coefficient satisfies the following relationship: In the formula, are the first and second fitting coefficients of the pth data processing segment, respectively; Using the pth data processing segment Q p Construct a fitting straight line segment to satisfy the following relationship: In the formula, is the pth data processing segment Q p The constructed fitting straight line segment, T p To fit a straight line segment The horizontal axis variable, To fit a straight line segment The ordinate variable, that is, in the pth data processing segment Q p Fitted straight line segment The horizontal axis value is T p The vertical coordinate value corresponding to time; and The pth data processing segment Q p Minimum and maximum power unit torque sampling data; S32: respectively process all P data processing segments {Q1, Q2, ..., Q p ,…,Q P }Construct P fitting straight line segments The P fitting straight line segments are spliced; the spliced ​​P fitting straight line segments are smoothed by using the cubic B-spline interpolation method to construct a torque-current mapping relationship curve that satisfies the following relationship: In the formula, represents the torque-current mapping relationship curve constructed by smoothing P fitting straight line segments, T is the torque-current mapping relationship curve The horizontal axis variable, y is the torque-current mapping curve The vertical axis variable, that is, when y is y=i d and y=i q When , formula (18) is written as and 6. The method for optimizing torque distribution of a multi-motor drive system according to claim 1, characterized in that: The S4 includes: S41: Let the current time be t, consider the time interval between the sampling time and the control variable output, predict the train speed with a step length of τ, and build a train speed prediction model that satisfies the following relationship: In the formula, v t+τ is the predicted train speed at time t+τ, v t is the train speed at time t, N is the total number of power units in the multi-motor drive system, mg is the positive pressure of the wheelset, τ is the prediction step, M is the mass of the train, r is the wheel diameter of the train, and f(·) represents the train resistance, which is a function of the train speed and mass. is the motor speed of the unit power unit at time t, is the adhesion coefficient of the unit power unit at time t, satisfying the following relationship: Where c1, c2, c3, c4 are the first, second, third, and fourth parameters of the power unit adhesion coefficient, unit = 1, 2, ..., N; S42: predicting the motor speed of the unit-th power unit of the multi-motor drive system with a step length of τ, and constructing a motor speed prediction model of the unit-th power unit of the multi-motor drive system, which satisfies the following relationship: In the formula, is the predicted motor speed of the unit power unit at time t+τ, η gear 、a gear Represent transmission efficiency and gear ratio respectively, J unit is the moment of inertia of the unit power unit, is the torque of the unit-th power unit at time t; S43: Predicting the total torque reference value of the multi-motor drive system with a step length of τ, and constructing a prediction model for the total torque reference value of the multi-motor drive system, which satisfies the following relationship: In the formula, is the predicted total torque reference value of the multi-motor drive system at time t+τ, Γ(·) represents the predicted train speed v at time t+τ t+τ The mapping relationship between the predicted total torque reference value of the multi-motor drive system at time t+τ is obtained from the traction characteristic curve of the train; S44: Predict the power loss of the unit power unit of the multi-motor drive system with a step size of τ, and use the established torque-current mapping curve to estimate the current d-axis and q-axis components of the unit power at different torques. Combined with the power loss model of the drive system power unit established in S1, a power loss prediction model for the unit power unit of the multi-motor drive system is constructed, which satisfies the following relationship: In the formula, is the predicted power loss of the unit-th power unit of the multi-motor drive system at time t+τ, f I,P (·,·,·,) represents the mapping relationship between the current and power loss of the unit-th power unit in the multi-motor drive system. and For y, they are y=i d and y=i q The torque-current mapping curve established when The torque reference value assigned to the unit power unit at time t+τ satisfies the following relationship: In the formula, is the distribution coefficient of the torque reference value of the unit power unit at time t+τ, is the predicted total torque reference value of the multi-motor drive system at time t+τ.

7. The method for optimizing torque distribution of a multi-motor drive system according to claim 1, characterized in that: The S5 includes: S51: According to the power loss prediction model of the unit power unit of the multi-motor drive system, an energy efficiency evaluation function of the multi-motor drive system is established to satisfy the following relationship: In the formula, represents the energy consumption evaluation function of the multi-motor drive system at the current moment, represents the power loss prediction model of the unit-th power unit of the multi-motor drive system; S52: According to the motor speed prediction model of the unit power unit of the multi-motor drive system and the wheel slip rate, a stability evaluation function of the multi-motor drive system is established to satisfy the following relationship: In the formula, represents the stability evaluation function of the multi-motor drive system at the current moment, It is expressed as the slip rate of the unit power unit at time t+τ, satisfying the following relationship: is the penalty coefficient of the slip rate of the unit power unit at time t+τ, satisfying the following relationship: In the formula, κ lim is the limit value of the power unit slip rate; S53: According to the allocation coefficient of the torque reference value of the unit power unit, a total torque following evaluation function of the multi-motor drive system is established to satisfy the following relationship: In the formula, is the total torque following evaluation function of the multi-motor drive system; S54: configuring weight coefficients for the three evaluation functions, establishing a system optimization objective function, and satisfying the following relationship: In the formula, ρ1, ρ2 and ρ3 are evaluation functions respectively. The weight coefficient of .

8. The method for optimizing torque distribution of a multi-motor drive system according to claim 1, characterized in that: The S6 includes: S61: Set the distribution coefficient of the torque reference value of each power unit at time t+τ to the control vector at the current time t The system optimization objective function is written as the control vector U t The functional form is denoted by J t =g(U t ), the system optimization objective function is in U t The gradient at is expressed as: In the formula, Indicates that the system optimization objective function is in U t The gradient at is the control vector U t The unit-th component in also represents the distribution coefficient of the unit-th power unit torque reference value at time t+τ; ▽ is the gradient operator; Indicates the control component of unit The gradient of is: In the formula, and Represent the current moment evaluation function and For the unit component at time t+τ The partial derivative of S62: According to the system optimization objective function gradient, the learning rate of the control vector is designed as follows: Design the learning rate of the unit-th control component to satisfy the following relationship: In the formula, is the unit-th control component at the current time t The learning rate, A lim is the gradient limit of each iteration of the control component, α lim is the maximum learning rate of the control component set; When the optimization objective function gradient satisfies When the unit-th control component is With the set step size Perform gradient descent; When the optimization objective function gradient satisfies When the unit-th control component is With the set learning rate α lim Perform gradient descent. unit=1,2,...,N, construct the learning rate vector of the system control vector at the current time t as represents the field of real numbers; S63: Set the initial solution of the control vector to: S64: According to the learning rate vector, the unit-th control component of the control vector Update to satisfy the following relationship: In the formula, and are the nth s +1st and nth s The unit-th control component in the control vector after iterations; N s,unit Iterate until the learning rate of the unit-th control component meets the search stop condition: α end is the learning rate that satisfies the stop search condition; the control component obtained at this time is the current time t through N s,unit The unit-th optimized control component after the iterative update That is, the optimal distribution coefficient of the torque reference value of the unit power unit at time t+τ; combined with the total reference torque of the system The optimized torque reference value allocated to the unit power unit is calculated as: unit=1,2,...,N, the optimal distribution coefficient of the torque reference values ​​of all N power units at the current time t is obtained as follows: The optimized torque reference value distributed by all N power units is 9. A torque optimization distribution system for a multi-motor drive system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images