A method for hybrid system configuration evaluation and parameter optimization

Abstract

The application discloses a kind of hybrid power system configuration evaluation and parameter optimization method, belong to hybrid power system technical field.This method obtains the system operating data under series, parallel and hybrid configuration, extracts power flow characteristic parameter and constructs design structure matrix;Spectral decomposition is carried out to design structure matrix, to determine the decoupling degree index, and the target configuration is selected accordingly;Under the target configuration, generate time series power sequence, perform resampling operation, set the parameter set to be optimized;Parameter evaluation matrix is constructed based on the parameter set to be optimized and design structure matrix, the adjustment direction of each parameter is determined by matrix decomposition, and the parameter set is updated iteratively along the direction, to obtain the optimized parameter set.This method uses spectral decomposition to realize the quantitative evaluation of configuration decoupling degree, determines the parameter adjustment direction by matrix decomposition, improves the objectivity of configuration selection and parameter optimization efficiency.

Description

Technical Field

[0001] This invention relates to the field of hybrid power system technology, specifically to a method for hybrid power system configuration evaluation and parameter optimization. Background Technology

[0002] Hybrid power systems can operate in various configurations, including series, parallel, and series-parallel configurations, with significant differences in power flow paths and energy management logic across these configurations. In the early stages of system development, a suitable configuration needs to be selected from the available options based on design requirements, and the operating points of the internal components within the selected configuration must be optimized and calibrated. Existing configuration evaluation methods are mostly based on steady-state efficiency analysis, often placing different configurations under the same standard operating condition and calculating their respective average system efficiencies using engine universal characteristic curves and motor efficiency graphs, then selecting the configuration with the higher efficiency. This approach only reflects the energy conversion effect under ideal steady-state conditions and cannot measure the coupling complexity of energy transfer within the configuration. In actual operation, power shunting and power cycling between the engine, motor, battery, and transmission mechanisms lead to system dynamic response lag and increased control dimensions. Relying solely on steady-state efficiency comparisons can easily overestimate the actual performance of some configurations with severe power flow circulation, resulting in selection bias. At the parameter optimization level, after determining the configuration, a global optimization method based on intelligent algorithms is commonly used. This treats control parameters such as engine torque and motor speed as independent variables, and calls the system simulation model for black-box iteration within preset boundaries. This optimization process is detached from the topological information of the configuration itself; the direction of parameter adjustment relies entirely on the feedback of the objective function. When the parameter space coupling is strong, the search path exhibits random walk characteristics, computational costs increase exponentially, and the optimization endpoint is prone to getting trapped in local maxima. Therefore, it is necessary to address the problem of the inability to quantitatively assess the degree of power coupling in the configuration, as well as the issues of the adjustment direction being disconnected from the structural characteristics of the configuration during parameter optimization, and slow convergence. Summary of the Invention

[0003] This paper provides a method for evaluating the configuration and optimizing the parameters of a hybrid power system, which solves the problems of high subjectivity in configuration selection due to the lack of objective quantitative means to measure the degree of power coupling and decoupling in the existing technology, and convergence lag caused by the failure of parameter optimization to utilize configuration structural information.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for evaluating and optimizing the configuration of a hybrid power system, comprising: acquiring system operation data of the hybrid power system under series, parallel, and series-parallel configurations respectively; extracting power flow characteristic parameters under each configuration based on the system operation data; constructing a design structure matrix based on the power flow characteristic parameters under each configuration, wherein the rows and columns of the design structure matrix correspond to multiple components of the hybrid power system, and the matrix elements represent the power transfer relationship between the corresponding components; performing spectral decomposition on the design structure matrix to obtain multiple eigenvalues ​​and corresponding eigenvectors; and determining the decoupling index of each configuration based on the multiple eigenvalues. The process involves: selecting a target configuration from the series, parallel, and hybrid configurations based on the decoupling index; generating a time-series power sequence based on the system operating data under the target configuration; performing a resampling operation on the time-series power sequence to obtain a resampled power sequence; setting a set of parameters to be optimized based on the resampled power sequence; constructing a parameter evaluation matrix based on the set of parameters to be optimized and the design structure matrix of the target configuration; performing matrix decomposition on the parameter evaluation matrix; determining the adjustment direction of each parameter to be optimized using the decomposition results; iteratively updating the set of parameters to be optimized according to the adjustment direction until a preset iteration termination condition is met to obtain the optimized parameter set.

[0005] As a technical solution of the present invention, obtaining system operation data of a hybrid power system under series, parallel, and series-parallel configurations respectively includes: the system operation data including engine speed, motor torque, battery state of charge, and bus voltage; controlling the hybrid power system test platform to operate in series configuration according to a preset first operating condition curve, and collecting engine speed, motor torque, battery state of charge, and bus voltage at a first sampling frequency during operation to obtain series configuration operation data; controlling the hybrid power system test platform to switch to parallel configuration, operating according to the first operating condition curve, and collecting engine speed, motor torque, battery state of charge, and bus voltage at the first sampling frequency to obtain parallel configuration operation data; controlling the hybrid power system test platform to switch to series-parallel configuration, operating according to the first operating condition curve, and collecting engine speed, motor torque, battery state of charge, and bus voltage at the first sampling frequency to obtain series-parallel configuration operation data; and summarizing the series configuration operation data, the parallel configuration operation data, and the series-parallel configuration operation data into the system operation data. Preferably, based on the system operating data, power flow characteristic parameters under each configuration are extracted, specifically including: the power flow characteristic parameters include the power shunt ratio and the power cycle coefficient; performing point-by-point multiplication on the engine speed and motor torque in the system operating data to obtain the real-time engine power and the real-time motor power; subtracting the real-time motor power from the real-time engine power to obtain the power difference, and dividing the power difference by the real-time engine power to obtain the power shunt ratio; calculating the battery charging and discharging power based on the rate of change of battery state of charge, using the ratio of the battery charging and discharging power to the real-time engine power as the first ratio, and performing an absolute value operation on the first ratio to obtain the power cycle coefficient.

[0006] As a technical solution of the present invention, a design structure matrix is ​​constructed based on the power flow characteristic parameters under each configuration. Specifically, this includes: determining the arrangement order of the multiple components according to the component topology of the hybrid power system, wherein the multiple components include an engine, a motor, a battery, and a transmission mechanism; mapping the power split ratio and power cycle coefficient in the power flow characteristic parameters to the initial values ​​of the matrix elements, wherein the mapping method is to assign the product of the power split ratio and the power cycle coefficient to the matrix elements; setting the values ​​of the matrix elements located on the main diagonal of the design structure matrix to zero; and retaining the initial values ​​of the matrix elements located outside the main diagonal of the design structure matrix, thereby obtaining the design structure matrix. Preferably, the design structure matrix is ​​subjected to spectral decomposition to obtain multiple eigenvalues ​​and corresponding eigenvectors. The decoupling index for each configuration is determined based on these eigenvalues. Specifically, this includes: performing eigenvalue decomposition on the design structure matrix to obtain multiple eigenvalues ​​and corresponding eigenvectors; calculating the difference between the number of positive eigenvalues ​​and the number of zero eigenvalues, dividing the difference by the order of the design structure matrix to obtain a first decoupling factor; calculating the ratio of the eigenvalue with the largest absolute value to the eigenvalue with the smallest absolute value to obtain a second decoupling factor; and using the product of the first and second decoupling factors as the decoupling index. By extracting power flow characteristic parameters and constructing the design structure matrix, combined with the decoupling index obtained from spectral decomposition, the complex coupling degree of energy flow in different hybrid power configurations can be accurately quantified. This provides an objective basis for configuration selection, avoids the limitations of judging the merits of configurations solely based on experience, and significantly improves the accuracy and efficiency of configuration evaluation.

[0007] As a technical solution of the present invention, under the target configuration, a time-series power sequence is generated based on the system operating data, and a resampling operation is performed on the time-series power sequence to obtain a resampled power sequence. Specifically, this includes: extracting target operating data corresponding to the target configuration from the system operating data; extracting engine power values ​​and motor power values ​​from the target operating data in chronological order to generate the time-series power sequence; uniformly resampling the time-series power sequence along the time axis, with a resampling time interval of a preset time step, to obtain an initial resampled sequence with uniform time intervals; and performing sliding window filtering on the initial resampled sequence, taking the median value within each window as the output value of that window, with a sliding step of one time step, to obtain the resampled power sequence. By performing resampling and sliding window filtering operations, high-frequency noise and instantaneous fluctuations in the original data are effectively removed, making the resampled power sequence more smoothly reflect the steady-state operating characteristics of the system, providing a high signal-to-noise ratio data foundation for subsequent parameter optimization, thereby improving the stability of the optimization results.

[0008] Preferably, based on the resampled power sequence, a set of parameters to be optimized is set, specifically including: the set of parameters to be optimized includes engine operating point parameters and motor operating point parameters; for each power value in the resampled power sequence, a preset engine universal characteristic data is searched to obtain the engine speed range and engine torque range corresponding to the power value; a first speed value is selected from the engine speed range as the engine operating point parameter, and a first torque value is selected from the engine torque range as the motor operating point parameter; the first speed value is substituted into the engine universal characteristic data to query the corresponding engine fuel consumption rate, and the first torque value is substituted into the preset motor efficiency characteristic data to query the corresponding motor efficiency value; speed values ​​with engine fuel consumption rates lower than a preset fuel consumption threshold are retained as candidate engine operating point parameters, and torque values ​​with motor efficiency values ​​higher than a preset efficiency threshold are retained as candidate motor operating point parameters, and the set of parameters to be optimized is composed of all candidate engine operating point parameters and candidate motor operating point parameters. This technical solution utilizes engine universal characteristic data and motor efficiency characteristic data to screen candidate operating points, concentrating the parameters to be optimized in the high-efficiency and low-energy-consumption regions. This effectively reduces the search space, accelerates the subsequent optimization convergence speed, and ensures that the final optimized parameters possess excellent fuel economy and system efficiency.

[0009] As a technical solution of the present invention, based on the set of parameters to be optimized and the design structure matrix of the target configuration, a parameter evaluation matrix is ​​constructed. Matrix decomposition is performed on the parameter evaluation matrix, and the adjustment direction of each parameter to be optimized is determined using the decomposition result. Specifically, this includes: using each parameter in the set of parameters to be optimized as a row index and each component in the design structure matrix of the target configuration as a column index to construct an initial parameter evaluation matrix, where the matrix elements are the sensitivity coefficients of the corresponding parameters to the corresponding components; performing orthogonal triangular decomposition on the initial parameter evaluation matrix to obtain an upper triangular matrix, and extracting the diagonal elements of the upper triangular matrix as the independent contribution of each parameter; sorting the independent contribution in descending order, marking the parameters corresponding to the independent contribution with odd-numbered sorting indices as positive adjustment parameters, and marking the parameters corresponding to the independent contribution with even-numbered sorting indices as negative adjustment parameters; and using the increasing direction of the positive adjustment parameters as the first adjustment direction and the decreasing direction of the negative adjustment parameters as the second adjustment direction. By performing orthogonal triangular decomposition on the parameter evaluation matrix, the interaction effects between parameters are effectively isolated, and the independent contribution of each parameter is extracted. Based on this, the parameters are adjusted in different directions, avoiding mutual interference between parameters that could cause oscillations in the optimization process, making the optimization path clearer and more efficient.

[0010] Preferably, the parameter set to be optimized is iteratively updated according to the adjustment direction until a preset iteration termination condition is met to obtain an optimized parameter set. Specifically, this includes: increasing all parameters in the first adjustment direction by a first preset step size, and decreasing all parameters in the second adjustment direction by a second preset step size to obtain an updated parameter set to be optimized; substituting the updated parameter set to be optimized into the system simulation model of the target configuration to calculate the overall system efficiency; comparing the overall system efficiency with the overall system efficiency of the previous iteration; if the overall system efficiency is improved, the current update is retained; if the overall system efficiency is decreased, the parameter set of the previous iteration is reverted to the parameter set of the previous iteration, and both the first preset step size and the second preset step size are reduced to half of their original values; when both the first preset step size and the second preset step size are less than a preset step size threshold, the preset iteration termination condition is met, and the current parameter set is used as the optimized parameter set. By using a backoff mechanism that dynamically reduces the step size, the search granularity is adaptively adjusted during the iterative optimization process. This ensures that the system can quickly approach the optimal region in the early stages and finely approximate the optimal solution in the later stages. It effectively balances global search capability and local convergence accuracy, and ultimately obtains the combination of engine and motor operating point parameters that maximizes the overall system efficiency.

[0011] In this invention, the preset engine universal characteristic data includes braking ratio fuel consumption data under different speed and torque combinations, and the preset motor efficiency characteristic data includes motor system efficiency data under different speed and torque combinations.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: By extracting the power split ratio and power cycle coefficient from the operating data of various hybrid power system configurations, these are mapped to elements of a design structure matrix. The rows and columns of this matrix correspond to components such as the engine, motor, battery, and transmission mechanism, and the elements express the power transfer relationships between these components. Eigenvalue decomposition is performed on the design structure matrix. The first decoupling factor is constructed using the ratio of the difference between the number of positive eigenvalues ​​and the number of zero eigenvalues ​​to the matrix order. Simultaneously, the ratio of the largest to the smallest absolute value of the eigenvalue is used as the second decoupling factor. Multiplying these two factors yields a decoupling degree index. This decoupling degree index simultaneously characterizes the number of dimensions and the strength level of power coupling in the system; a higher value indicates more independent power flow between components and fewer coupling loops. Based on this index, target configurations are selected from series, parallel, and hybrid configurations. This avoids the kinetic coupling costs ignored by comparing only steady-state efficiency, providing clear quantitative criteria for selection decisions, reducing the workload of repeated modeling and simulation experiments, and making the selected configuration easier to design for subsequent decoupling control and energy management strategies. Under the target configuration, engine power and motor power values ​​are extracted from the corresponding system operation data in a time sequence. After uniform resampling and sliding window mid-range filtering, a resampled power sequence with transient pulsations removed is generated. Based on this sequence, candidate engine and motor operating point parameters are obtained through table lookup and threshold filtering, forming a set of parameters to be optimized. This set of parameters is combined with the design structure matrix of the target configuration to construct a parameter evaluation matrix, where each element is assigned a sensitivity coefficient of the corresponding parameter to the corresponding component. The parameter evaluation matrix is ​​decomposed by orthogonal trigonometric decomposition, and the diagonal elements of the upper triangular matrix are extracted to obtain the independent contribution of each parameter, independent of the effects of other parameters. The independent contributions are sorted in descending order, and parameters are marked as positive or negative adjustments based on their parity, thus establishing the direction of increase or decrease for each parameter to be optimized. This process maps the power flow coupling characteristics within the configuration to the parameter sensitivity space. The determined adjustment direction matches the intrinsic structure of the design structure matrix, ensuring that each iteration step proceeds along the decoupling direction of the parameter space, effectively suppressing competition and interference between variables. During iterative parameter updates, positive adjustments increase parameters, while negative adjustments decrease them. The step size is scaled based on changes in overall system efficiency, and the optimization process terminates when the step size decays below a threshold. Compared to traditional unstructured global search schemes, the resulting optimized parameter set achieves higher overall system efficiency with fewer iterations. Furthermore, the parameter combinations better align with the inherent decoupling characteristics of the configuration, reducing repeated debugging during actual calibration, shortening the development cycle, and improving the overall energy utilization rate of the hybrid power system. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a flowchart of the hybrid power system configuration evaluation and parameter optimization method; Figure 2 This is a flowchart of the hybrid power system configuration evaluation and parameter optimization method; Figure 3 These are the time-series curves showing the power split ratio variation of different configurations of a hybrid power system. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] See Figure 1 This invention provides a method for evaluating and optimizing the configuration of a hybrid power system, comprising: acquiring system operation data of the hybrid power system under series, parallel, and series-parallel configurations respectively; extracting power flow characteristic parameters under each configuration based on the system operation data; constructing a design structure matrix based on the power flow characteristic parameters under each configuration, wherein the rows and columns of the design structure matrix correspond to multiple components of the hybrid power system, and the matrix elements represent the power transfer relationship between the corresponding components; performing spectral decomposition on the design structure matrix to obtain multiple eigenvalues ​​and corresponding eigenvectors; and determining the decoupling index of each configuration based on the multiple eigenvalues. The process involves: selecting a target configuration from series, parallel, and hybrid configurations based on the decoupling index; generating a time-series power sequence based on system operating data under the target configuration; performing a resampling operation on the time-series power sequence to obtain a resampled power sequence; defining a set of parameters to be optimized based on the resampled power sequence; constructing a parameter evaluation matrix based on the parameter set and the design structure matrix of the target configuration; performing matrix decomposition on the parameter evaluation matrix; and using the decomposition results to determine the adjustment direction of each parameter to be optimized; iteratively updating the parameter set according to the adjustment direction until a preset iteration termination condition is met to obtain the optimized parameter set.

[0017] In specific implementation, please refer to Figure 2The hybrid power system configuration evaluation and parameter optimization method includes the process of acquiring system operating data and extracting power flow characteristic parameters.

[0018] The system operation data of the hybrid power system is acquired under series, parallel, and series-parallel configurations. This data includes engine speed, motor torque, battery state of charge (SBC), and bus voltage. The hybrid power system test platform comprises an engine, motor, battery, transmission mechanism, and a power coupling device with switchable configurations. The test platform is controlled to operate in series configuration according to a preset first operating condition curve. This first operating condition curve is a time-vehicle speed sequence pre-stored in the test platform control unit, covering idle, acceleration, constant speed, and deceleration phases. During operation, engine speed, motor torque, battery SBC, and bus voltage are collected at a first sampling frequency of 100 Hz, based on the signal bandwidth, to fully capture engine speed fluctuations and transient changes in motor torque. Series configuration operation data is also acquired. Each sampling point in the series configuration operation data includes a timestamp, engine speed value, motor torque value, battery SBC value, and bus voltage value.

[0019] The hybrid power system test platform is switched to a parallel configuration, and the power coupling device is adjusted to a parallel connection between the engine and the motor. It then operates again according to the first operating condition curve, collecting engine speed, motor torque, battery state of charge, and bus voltage at the first sampling frequency to obtain parallel configuration operating data. The hybrid power system test platform is then switched to a series-parallel configuration, and the power coupling device is adjusted to a power-split state. It operates according to the first operating condition curve, collecting engine speed, motor torque, battery state of charge, and bus voltage at the first sampling frequency to obtain series-parallel configuration operating data. The series, parallel, and series-parallel configuration operating data are aggregated into system operating data and stored according to their respective configurations.

[0020] Power flow characteristic parameters for each configuration are extracted based on system operating data. These parameters include the power split ratio and the power cycle coefficient. For a specific configuration's system operating data, the following processing is performed: A point-by-point multiplication operation is performed on the engine speed and motor torque in the system operating data. For each sampling time, the engine speed value is multiplied by the motor torque value to obtain the real-time engine power, and the motor torque value is multiplied by the motor speed value to obtain the real-time motor power. The motor speed can be obtained through feedback from the motor controller or calculated from the vehicle speed and transmission ratio. The time series sequences of the engine real-time power and the motor real-time power are obtained through the point-by-point multiplication operation.

[0021] Calculate the power shunt ratio: Subtract the real-time motor power from the real-time engine power at each sampling moment to obtain the power difference. Divide the power difference by the real-time engine power to obtain the power shunt ratio. Calculate the power cycle factor: Calculate the battery charging and discharging power based on the rate of change of battery state of charge. The formula for calculating battery charging and discharging power is: ; in, This indicates the battery's charging and discharging power, measured in watts. It represents the difference in battery state of charge between two adjacent sampling times, which is obtained by subtracting the battery state of charge from the battery state of charge at the previous time from the battery state of charge at the later time. It represents the time interval between adjacent sampling moments, which is equal to the reciprocal of the first sampling frequency; This indicates the battery's rated capacity, measured in ampere-hours (AH), and is provided on the battery's nameplate. This indicates the battery's rated voltage, measured in volts, and is determined by the battery specifications. (Calculated...) Afterwards, The ratio of the power output to the engine's real-time power output at the same moment is used as the first ratio. An absolute value operation is performed on the first ratio to obtain the power cycle coefficient.

[0022] The above extraction process is performed on the operating data of the series configuration, the parallel configuration, and the mixed configuration to obtain the power flow characteristic parameters corresponding to each configuration. The power flow characteristic parameters include the power shunting ratio and the power cycling coefficient at each sampling time in the configuration.

[0023] See Figure 3 In the figure, the horizontal axis represents the sampling point number, ranging from 0 to 500. This represents the data sequence of 500 consecutive sampling points collected during the operation of the hybrid power system test platform according to the preset first operating condition curve, for the series, parallel, and series-parallel configurations respectively. The vertical axis represents the power split ratio, which ranges from 0 to 1, reflecting the power distribution relationship between the engine and the motor in the hybrid power system.

[0024] The three curves in the figure correspond to the series configuration (blue solid line, marked with dots), the parallel configuration (orange dashed line, marked with squares), and the mixed configuration (green dotted line, marked with triangles), respectively. The shaded area around the curves represents the fluctuation range of the power split ratio, reflecting the discreteness of the sampling data and the measurement error.

[0025] The power split ratio of the series configuration remains consistently high, generally ranging from 0.85 to 0.99, exhibiting slight periodic fluctuations. This indicates a higher proportion of engine power and a smaller proportion of motor power contribution, consistent with the characteristic of the engine as the primary power source in a series configuration. The power split ratio of the hybrid configuration remains between approximately 0.50 and 0.60, with relatively stable fluctuations, showing a more balanced power distribution between the engine and motor, reflecting the characteristics of power splitting in a hybrid configuration. The power split ratio of the parallel configuration is low, approximately between 0.10 and 0.20, with a clear fluctuation pattern, indicating that the motor dominates the power output, while the engine power proportion is relatively low, consistent with the actual power distribution pattern of the motor and engine operating in parallel in a parallel configuration.

[0026] In practical implementation, a design structure matrix is ​​constructed based on the power flow characteristic parameters under each configuration. The rows and columns of the design structure matrix correspond to multiple components of the hybrid power system. The components of the hybrid power system include the engine, motor, battery, and transmission mechanism. When determining the arrangement order of multiple components, based on the topological relationship of the hybrid power system components, that is, according to the connection order between components in the process of power generation and transmission to the wheels, the engine is placed in the first position, the motor in the second position, the battery in the third position, and the transmission mechanism in the fourth position, forming a fourth-order square matrix structure.

[0027] The power shunting ratio and power cycling coefficient in the power flow characteristic parameters are mapped to the initial values ​​of matrix elements. For any matrix element located in row u and column v of the design structure matrix, the mapping method is to assign the product of the power shunting ratio and the power cycling coefficient to the matrix element. Row u corresponds to the output power component, and column v corresponds to the input power component. The power shunting ratio is taken as the average value of the power shunting ratio at all sampling times under the corresponding configuration, and the power cycling coefficient is taken as the average value of the power cycling coefficient at all sampling times under the corresponding configuration. Taking the element in the row corresponding to the engine and the column corresponding to the motor as an example, the mapped value of this element is the product of the average power shunting ratio and the average power cycling coefficient.

[0028] For the matrix element located on the main diagonal of the design structure matrix, i.e., the element in the u-th row and u-th column, its value is set to zero. The main diagonal element represents the power transfer from the component itself, and does not generate external interaction in the power flow transfer relationship. For the matrix elements located outside the main diagonal of the design structure matrix, their initial values ​​are retained, resulting in the design structure matrix. The engine, motor, battery, and transmission mechanism are arranged in a 4x4 matrix according to their order. The above construction process is performed on the series configuration, parallel configuration, and hybrid configuration respectively to obtain the design structure matrix corresponding to each configuration.

[0029] Spectral decomposition of the design structure matrix yields multiple eigenvalues ​​and their corresponding eigenvectors. The eigenvalue decomposition operation is performed on the design structure matrix, denoted as [Eigenvalue decomposition]. The matrix has an order n of 4. After performing eigenvalue decomposition, n eigenvalues ​​λ are obtained. a1 ,λ a2 ,...,λ an and the corresponding n feature vectors v a1 ,v a2 ,...,v an , where 'a' is an identifier to distinguish different configurations: a=1 for series configuration, a=2 for parallel configuration, and a=3 for mixed configuration.

[0030] Determine the decoupling index for each configuration based on multiple eigenvalues. Calculate the difference between the number of positive eigenvalues ​​and the number of zero eigenvalues. Traverse the eigenvalue sequence; for the b-th eigenvalue... ,like The value is greater than 1×10 -6 Then it is determined to be a positive eigenvalue, if The absolute value is less than or equal to 1 × 10 -6 Then it is determined to be a zero eigenvalue. The number of positive eigenvalues ​​is denoted as . The number of zero eigenvalues ​​is denoted as Calculate the difference The difference Divide by n to obtain the first decoupling factor. The first decoupling factor has a value range between -1 and 1.

[0031] Calculate the ratio of the eigenvalue with the largest absolute value to the eigenvalue with the smallest absolute value among a set of eigenvalues. Traverse the sequence of eigenvalues ​​and find the eigenvalue with the largest absolute value, denoted as . Find the eigenvalue with the smallest absolute value that is not zero; its absolute value is denoted as . .like If the absolute value is zero, then take the second smallest non-zero absolute value as... Calculate the second decoupling factor. The value of the second decoupling factor is greater than or equal to 1. The product of the first and second decoupling factors is used as the decoupling degree index, and the formula for calculating the decoupling degree index is as follows: ; in, The decoupling index represents the degree of coupling for the a-th configuration; The first decoupling factor is obtained by dividing the difference between the number of positive eigenvalues ​​and the number of zero eigenvalues ​​by the matrix order. The second decoupling factor is obtained by the ratio of the eigenvalue with the largest absolute value to the eigenvalue with the smallest absolute value. Represents the number of positive eigenvalues Number of zero eigenvalues The difference; n represents the order of the design structure matrix, which takes the value of 4; This represents the absolute value of the eigenvalue with the largest absolute value. It represents the absolute value of the eigenvalue that is the smallest and not zero.

[0032] Calculate the decoupling index of the tandem configuration according to the above process. Decoupling index of parallel configuration Decoupling index of hybrid configuration .Compare , and The configuration with the highest decoupling index value is selected as the target configuration. A higher decoupling index indicates lower power coupling between components in the configuration, resulting in higher power flow controllability. The identifier of the target configuration and its corresponding design structure matrix are stored for subsequent parameter optimization.

[0033] In practice, after the target configuration is determined, a time-series power sequence is generated based on the system operating data, and a resampling operation is performed on the time-series power sequence. Target operating data corresponding to the target configuration is extracted from the system operating data. The target configuration is the one with the highest decoupling index value among series, parallel, and hybrid configurations; the target operating data is the operating data corresponding to this configuration. The target operating data contains multiple sampling points arranged in chronological order. Each sampling point includes a timestamp, engine speed value, motor torque value, battery state of charge value, and bus voltage value.

[0034] Engine power and motor power values ​​are extracted from the target operating data in chronological order to generate a time-series power sequence. The engine power value is obtained by multiplying the engine speed and engine torque value at each sampling point in the target operating data. The engine torque value is obtained from the torque parameter fed back by the engine control unit. The motor power value is obtained by multiplying the motor torque and motor speed value at each sampling point in the target operating data. The motor speed value is obtained from the speed parameter fed back by the motor controller. The engine power and motor power values ​​corresponding to each sampling point are arranged in chronological order to form a time-series power sequence, where each time point corresponds to one engine power value and one motor power value.

[0035] The time-series power sequence is uniformly resampled along the time axis, with a pre-defined time step interval. The pre-defined time step is set to 1 second, based on the fact that the system dynamic response time constant of the target configuration is on the order of seconds; a 1-second time step can smooth transient fluctuations and preserve power change trends. Using the start time of the time-series power sequence as the base time, a resampled time point sequence is generated on the time axis at pre-defined time step intervals. For each resampled time point, the two original sampling points with the closest timestamps in the time-series power sequence are found. The engine power value and motor power value corresponding to the resampled time point are calculated through linear interpolation, resulting in an initial resampled sequence with uniform time intervals. The time interval between adjacent sampling points in the initial resampled sequence is also the pre-defined time step.

[0036] A sliding window filter is applied to the initial resampled sequence. The sliding window length is set to 5 sampling points, based on the principle that 5 sampling points cover a 5-second time span with a 1-second time step, effectively filtering out short-term power spikes without overly smoothing the power curve. The sliding step is set to one time step, i.e., 1 second. The sliding window starts from the first sampling point of the initial resampled sequence and slides one sampling point forward each time. At each window position, the engine power values ​​of all sampling points within the window are extracted, sorted by numerical value, and the median value is taken as the median engine power value for that window position. The motor power values ​​of all sampling points within the window are extracted, sorted by numerical value, and the median value is taken as the median motor power value for that window position. The median engine power value and the median motor power value are combined as the output value for that window position. After the sliding window traverses the entire initial resampled sequence, a resampled power sequence consisting of the output values ​​for each window position is obtained. Each sampling point in the resampled power sequence contains the median engine power and the median motor power after sliding window filtering.

[0037] Window length of sliding window filtering The calculation formula is: ; in, This indicates the number of sampling points contained in the sliding window, with a value of 5. This indicates the filtering time span, set to 2 seconds. The basis for this setting is that the power response rise time of the hybrid power system is within 2 seconds, and a 2-second time span can effectively cover a single power adjustment process. This indicates the preset time step for resampling, with a value of 1 second. This indicates the floor function. The calculation result is The number of sampling points obtained for the sliding window is 5. Using an odd window length ensures that each window contains a unique median.

[0038] In practice, a set of parameters to be optimized is set based on the resampled power sequence. This set includes engine operating point parameters and motor operating point parameters. The engine operating point parameter is the engine speed value, and the motor operating point parameter is the motor torque value. The resampled power sequence contains multiple sampling points, each with the median engine power and median motor power after sliding window filtering.

[0039] Perform the following processing procedure on each power value in the resampled power sequence. The power value refers to the median engine power at each sampling point in the resampled power sequence. Locate the preset engine universal characteristic data, which includes brake-to-fuel consumption data under different speed and torque combinations. The engine universal characteristic data is stored in a two-dimensional table. Rows in the two-dimensional table correspond to engine speed values, with row indices ranging from 800 rpm to 6000 rpm at 100 rpm intervals. Columns in the two-dimensional table correspond to engine torque values, with column indices ranging from 10 Nm to the maximum torque value at 10 Nm intervals. Each cell in the two-dimensional table stores the brake-to-fuel consumption value for the corresponding speed and torque combination, with the unit being grams per kilowatt-hour.

[0040] For a median engine power value in the resampled power sequence, the corresponding engine speed range and engine torque range are searched in the engine universal characteristic data. Engine power is the product of speed and torque. In the two-dimensional table of engine universal characteristic data, all speed and torque combinations are iterated, and the power value corresponding to each combination is calculated. Speed ​​and torque combinations whose power values ​​fall within the upper and lower deviation ranges of the engine power median are selected, with the deviation range set to ±5% of the engine power median. The selected speed values ​​constitute the engine speed range, with the lower limit being the selected minimum speed value and the upper limit being the selected maximum speed value. Similarly, the selected torque values ​​constitute the engine torque range, with the lower limit being the selected minimum torque value and the upper limit being the selected maximum torque value.

[0041] A first engine speed value is selected from the engine speed range as the engine operating point parameter. This selection method involves taking the arithmetic mean of the lower and upper limits of the engine speed range, and using this arithmetic mean as the first engine speed value. Similarly, a first torque value is selected from the engine torque range as the motor operating point parameter. This selection method involves taking the arithmetic mean of the lower and upper limits of the engine torque range, and using this arithmetic mean as the first torque value. The first engine speed value is then substituted into the engine universal characteristic data to query the corresponding engine fuel consumption rate. This query is performed by finding the cell in the two-dimensional table row corresponding to the first engine speed value that is closest to the torque value corresponding to the current engine power median, and extracting the braking ratio fuel consumption value of that cell as the engine fuel consumption rate. Finally, the first torque value is substituted into preset motor efficiency characteristic data to query the corresponding motor efficiency value. This preset motor efficiency characteristic data includes motor system efficiency data under different speed and torque combinations. Motor efficiency characteristic data is stored in a two-dimensional table. Rows correspond to motor speed values, ranging from 0 to 12,000 revolutions per minute (rpm), distributed at 200 rpm intervals. Columns correspond to motor torque values, ranging from 0 Nm to the maximum torque value, distributed at 5 Nm intervals. Each cell in the table stores the motor system efficiency value for the corresponding speed and torque combination, expressed as a percentage. The query method involves calculating the ratio of the median motor power to the first torque value at the current sampling point as an estimated motor speed value. Then, the row corresponding to the estimated motor speed value and the column corresponding to the first torque value are located in the motor efficiency characteristic data. The motor system efficiency value at the intersection of these two values ​​is extracted as the final motor efficiency value.

[0042] Engine speeds with fuel consumption rates below a preset fuel consumption threshold are retained as candidate engine operating point parameters. The preset fuel consumption threshold is set as the minimum of all braking ratio fuel consumption values ​​in the engine universal characteristic data multiplied by 1.05. This threshold is based on selecting operating points with fuel consumption values ​​within 5% of the optimal level as candidates, ensuring that the 5% margin covers a reasonable range of the engine's efficient operating area. Torque values ​​with motor efficiency values ​​above a preset efficiency threshold are retained as candidate motor operating point parameters. The preset efficiency threshold is set as the maximum of all motor system efficiency values ​​in the motor efficiency characteristic data multiplied by 0.95. This threshold is based on selecting operating points with efficiency values ​​within 5% of the highest level as candidates, ensuring that the 5% margin covers a reasonable range of the motor's efficient operating area. The above process is repeated for the median engine power at each sampling point in the resampled power sequence. All engine speeds satisfying a fuel consumption rate below the preset fuel consumption threshold are aggregated to form a candidate engine operating point parameter set; all torque values ​​satisfying a motor efficiency value above the preset efficiency threshold are aggregated to form a candidate motor operating point parameter set. The set of parameters to be optimized consists of all candidate engine operating point parameters and all candidate motor operating point parameters.

[0043] The size of the candidate engine operating point parameter set and the candidate motor operating point parameter set is determined by the length of the resampled power sequence and the power distribution range. The parameter set to be optimized is stored in vector form, with the first half of the vector storing the candidate engine operating point parameters and the second half storing the candidate motor operating point parameters. The total length of the vector is... The calculation formula is: ; in, This represents the total number of parameters in the parameter set to be optimized; The number of candidate engine operating point parameters is obtained by accumulating the number of speed values ​​that satisfy the engine fuel consumption rate being lower than the preset fuel consumption threshold. The number of candidate motor operating point parameters is calculated by summing the number of torque values ​​that satisfy a motor efficiency value higher than a preset efficiency threshold. Each candidate motor operating point parameter in the parameter set to be optimized is assigned an independent index, with index values ​​numbered consecutively from 1 to... .

[0044] In practice, a parameter evaluation matrix is ​​constructed based on the design structure matrix of the parameter set to be optimized and the target configuration, and matrix decomposition is performed on the parameter evaluation matrix. The adjustment direction of each parameter to be optimized is determined by using the decomposition results.

[0045] Each parameter in the parameter set to be optimized is used as a row index. The parameter set to be optimized includes candidate engine operating point parameters and candidate motor operating point parameters, with a total number of parameters. Each component in the design structure matrix of the target configuration is used as a column index. The design structure matrix of the target configuration is a 4x4 square matrix, and the components include the engine, motor, battery, and transmission mechanism, with 4 columns. An initial parameter evaluation matrix is ​​then constructed. The initial parameter evaluation matrix has 10 rows. The initial parameter evaluation matrix has 4 columns. The row index r of the initial parameter evaluation matrix corresponds to the r-th parameter in the parameter set to be optimized, and the column index c of the initial parameter evaluation matrix corresponds to the c-th component in the design structure matrix of the target configuration. c=1 represents the engine, c=2 represents the motor, c=3 represents the battery, and c=4 represents the transmission mechanism.

[0046] The elements of the initial parameter evaluation matrix are the sensitivity coefficients of the corresponding parameters to the corresponding components. The sensitivity coefficients are calculated as follows: For the r-th parameter in the parameter set to be optimized, if the r-th parameter is a candidate engine operating point parameter, then in the system simulation model of the target configuration, the value of the r-th parameter is increased by 1%, keeping the other parameters unchanged. The system simulation model is run, and the power output change corresponding to the c-th component is recorded. The ratio of the power output change to the value change of the r-th parameter is used as the sensitivity coefficient. If the r-th parameter is a candidate motor operating point parameter, then the r-th parameter is increased by 1% in the system simulation model, the power output change of the c-th component is recorded, and the same ratio is calculated to obtain the sensitivity coefficient. This calculation is performed for all combinations of r and c, completing the assignment of values ​​to all elements of the initial parameter evaluation matrix.

[0047] The initial parameter evaluation matrix is ​​decomposed into an orthogonal triangular matrix using a modified Gram-Schmidt orthogonalization algorithm. and an upper triangular matrix The product of, i.e. Upper triangular matrix The number of rows and columns is 4, and the matrix elements are... The value is zero when the row index p is greater than the column index q; the non-zero value obtained from the decomposition is stored when the row index p is less than or equal to the column index q. Extracting the upper triangular matrix. The diagonal elements represent the independent contributions of each parameter; the diagonal elements are... , , and Each diagonal element corresponds to the independent contribution value of the component represented by a column index.

[0048] The independent contributions are sorted in descending order, with the four diagonal elements ordered from largest to smallest, resulting in an independent contribution sequence of 1, 2, 3, and 4. Parameters corresponding to independent contributions with odd-numbered sequences are marked as positive adjustment parameters; that is, parameters associated with components corresponding to sequences 1 and 3 are marked as positive adjustment parameters. Positive adjustment parameters are parameters in the parameter set to be optimized that have a row index correspondence with the components corresponding to sequences 1 and 3. Parameters corresponding to independent contributions with even-numbered sequences are marked as negative adjustment parameters; that is, parameters associated with components corresponding to sequences 2 and 4 are marked as negative adjustment parameters. Negative adjustment parameters are parameters in the parameter set to be optimized that have a row index correspondence with the components corresponding to sequences 2 and 4.

[0049] The direction of increase for positive adjustment parameters is defined as the first adjustment direction, and the direction of decrease for negative adjustment parameters is defined as the second adjustment direction. The first adjustment direction means that during the parameter update process, the values ​​of all parameters belonging to the positive adjustment parameter category are adjusted in the direction of increase; the second adjustment direction means that during the parameter update process, the values ​​of all parameters belonging to the negative adjustment parameter category are adjusted in the direction of decrease.

[0050] Iteratively update the parameter set to be optimized according to the adjustment direction to obtain the optimized parameter set. Before the iteration begins, initialize the first preset step size. The second preset step size is initialized to one-tenth of the standard deviation of the operating point parameter values ​​of all candidate engines in the parameter set to be optimized. The initial value of the first preset step size is one-tenth of the standard deviation of the operating point parameter values ​​of all candidate motors in the parameter set to be optimized. The initial value of the first preset step size is set based on the following: one-tenth of the standard deviation can produce a distinguishable change in the overall system efficiency in the parameter space, while avoiding the step size being too large, which would cause the iteration process to skip the optimal parameter region.

[0051] In each iteration, all parameters in the first adjustment direction are increased by a first preset step size, that is, the value of each candidate engine operating point parameter or candidate motor operating point parameter belonging to the positive adjustment parameter is added by the first preset step size; all parameters in the second adjustment direction are decreased by a second preset step size, that is, the value of each candidate engine operating point parameter or candidate motor operating point parameter belonging to the negative adjustment parameter is subtracted by the second preset step size. After performing the addition and subtraction operations, an updated set of parameters to be optimized is obtained.

[0052] The updated set of parameters to be optimized is substituted into the system simulation model of the target configuration to calculate the overall system efficiency. The system simulation model is a power flow calculation model established based on the component topology of the target configuration. The inputs are engine speed, motor torque, initial battery state of charge, and operating condition curves. The output is the overall system efficiency value over the entire operating cycle. The overall system efficiency is the ratio of the sum of mechanical energy output from the drive wheels to the sum of chemical energy consumed by the fuel during the operating cycle.

[0053] The overall system efficiency calculated in this iteration is compared with that of the previous iteration. If the overall system efficiency is greater than that of the previous iteration, the overall system efficiency is considered to have improved, the update is retained, and the updated set of parameters to be optimized is used as the current parameter set for the next iteration. If the overall system efficiency is less than or equal to that of the previous iteration, the overall system efficiency is considered to have decreased, and the system reverts to the parameter set of the previous iteration. This means discarding the updated set of parameters to be optimized, restoring the parameter set values ​​from the previous iteration, and reducing both the first and second preset step sizes to half their original values.

[0054] The preset step size threshold is set to five times the minimum resolution of the parameter values ​​in the parameter set to be optimized. The minimum resolution of the engine speed value in the parameter set to be optimized is 10 revolutions per minute, and the minimum resolution of the motor torque value is 1 N·m. The preset step size threshold is set to... For engine speed parameters, The value is taken as 10 × 5 = 50 revolutions per minute; for motor torque parameters, The value is 1 × 5 = 5 N·m. The iteration termination condition is determined as follows: when the first preset step size... Second preset step size All are less than the preset step size threshold When the preset iteration termination condition is met, the first preset step size is reached. Less than 50 and the second preset step size When the value is less than 5, the iteration process terminates.

[0055] When the iteration termination condition is met, the current parameter set is used as the optimized parameter set. The optimized parameter set includes the candidate engine operating point parameter values ​​and candidate motor operating point parameter values ​​that have been finally determined after iterative updates. The optimized parameter set is stored as the optimal operating point configuration parameters for the target configuration under the corresponding operating conditions.

[0056] The rationale for reducing the step size during iteration is as follows: when updating the parameter direction does not lead to an overall improvement in system efficiency, it indicates that the current step size exceeds the neighborhood where efficiency changes monotonically in the parameter space. Reducing the step size to half allows for a re-search for directions to improve efficiency within a smaller neighborhood. The coefficient for the step size reduction operation is set to half because the bisection method has a linear convergence speed in the search for extrema of continuous functions, which can balance search accuracy and the number of iterations.

[0057] During the iteration process, the calculated value of the overall system efficiency for each iteration, along with the corresponding first and second preset step size values, are recorded in the iteration log. The iteration log stores the iteration number, the overall system efficiency value, the first preset step size value, the second preset step size value, and the values ​​of all parameters in the updated set of parameters to be optimized for each iteration.

[0058] The convergence formula for the iterative process is: ; in, Represents the iteration convergence coefficient; This represents the overall system efficiency value obtained from the current iteration. This represents the overall system efficiency value obtained from the previous iteration. This represents the absolute value operation. When... When the convergence coefficient is less than 0.001 for three consecutive iterations, the overall system efficiency is considered to have stabilized. The step size threshold condition is then used to determine whether to terminate the iteration. The convergence coefficient threshold of 0.001 is set based on the following: when the relative change in efficiency is less than one-thousandth, the efficiency improvement brought about by parameter adjustment has entered the measurement error range, and further iteration no longer makes a significant contribution to the actual improvement of the overall system efficiency.

[0059] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for evaluating the configuration and optimizing the parameters of a hybrid power system, characterized in that, include: The system operation data of the hybrid power system under series configuration, parallel configuration and series-parallel configuration are obtained respectively, and the power flow characteristic parameters under each configuration are extracted based on the system operation data; Based on the power flow characteristic parameters under each configuration, a design structure matrix is ​​constructed. The rows and columns of the design structure matrix correspond to multiple components of the hybrid power system, and the matrix elements represent the power transfer relationship between the corresponding components. The design structure matrix is ​​subjected to spectral decomposition to obtain multiple eigenvalues ​​and corresponding eigenvectors. The decoupling index of each configuration is determined based on the multiple eigenvalues. Based on the decoupling index, a target configuration is selected from the series configuration, parallel configuration, and hybrid configuration; Under the target configuration, a time-series power sequence is generated based on the system operating data, and a resampling operation is performed on the time-series power sequence to obtain a resampled power sequence; Based on the resampled power sequence, a set of parameters to be optimized is set. Based on the set of parameters to be optimized and the design structure matrix of the target configuration, a parameter evaluation matrix is ​​constructed. Matrix decomposition is performed on the parameter evaluation matrix, and the adjustment direction of each parameter to be optimized is determined using the decomposition results. According to the adjustment direction, the set of parameters to be optimized is iteratively updated until the preset iteration termination condition is met, and the optimized set of parameters is obtained.

2. The hybrid power system configuration evaluation and parameter optimization method according to claim 1, characterized in that, Acquire system operation data of the hybrid power system under series, parallel, and series-parallel configurations, specifically including: The system operating data includes engine speed, motor torque, battery state of charge, and bus voltage. The test platform for the control hybrid power system operates in series configuration according to a preset first operating condition curve. During operation, engine speed, motor torque, battery state of charge and bus voltage are collected at the first sampling frequency to obtain series configuration operation data. The hybrid power system test platform is controlled to switch to parallel configuration and run according to the first operating condition curve. The engine speed, motor torque, battery state of charge and bus voltage are collected according to the first sampling frequency to obtain parallel configuration operation data. The hybrid power system test platform is controlled to switch to a series-parallel configuration and run according to the first operating condition curve. The engine speed, motor torque, battery state of charge and bus voltage are collected according to the first sampling frequency to obtain the series-parallel configuration operation data. The system operation data is aggregated from the series configuration operation data, the parallel configuration operation data, and the hybrid configuration operation data.

3. The hybrid power system configuration evaluation and parameter optimization method according to claim 2, characterized in that, Based on the system operating data, power flow characteristic parameters for each configuration are extracted, specifically including: The power flow characteristic parameters include the power split ratio and the power circulation coefficient; Perform a point-by-point multiplication operation on the engine speed and motor torque in the system operation data to obtain the real-time engine power and real-time motor power; The power difference is obtained by subtracting the real-time power of the motor from the real-time power of the engine, and the power split ratio is obtained by dividing the power difference by the real-time power of the engine. The battery charging and discharging power is calculated based on the rate of change of the battery's state of charge. The ratio of the battery charging and discharging power to the engine's real-time power is used as the first ratio. An absolute value operation is performed on the first ratio to obtain the power cycle coefficient.

4. The hybrid power system configuration evaluation and parameter optimization method according to claim 1, characterized in that, Based on the power flow characteristic parameters under each configuration, a design structure matrix is ​​constructed, specifically including: Based on the component topology of the hybrid power system, the arrangement order of the multiple components is determined, including the engine, motor, battery, and transmission mechanism. The power shunt ratio and power circulation coefficient in the power flow characteristic parameters are mapped to the initial values ​​of matrix elements. The mapping method is to assign the product of the power shunt ratio and the power circulation coefficient to the matrix elements. For the matrix elements located on the main diagonal of the design structure matrix, set their values ​​to zero; For the matrix elements located outside the main diagonal in the design structure matrix, retain their initial values ​​to obtain the design structure matrix.

5. The hybrid power system configuration evaluation and parameter optimization method according to claim 1, characterized in that, The design structure matrix is ​​subjected to spectral decomposition to obtain multiple eigenvalues ​​and corresponding eigenvectors. Based on these eigenvalues, a decoupling index for each configuration is determined, specifically including: Perform eigenvalue decomposition on the design structure matrix to obtain multiple eigenvalues ​​λ1, λ2, ..., λ n and the corresponding feature vectors v1, v2, ..., v n , where n is the order of the design structure matrix; Calculate the difference between the number of positive eigenvalues ​​and the number of zero eigenvalues ​​among the plurality of eigenvalues, and divide the difference by n to obtain the first decoupling factor; The ratio of the eigenvalue with the largest absolute value to the eigenvalue with the smallest absolute value among the plurality of eigenvalues ​​is calculated to obtain the second decoupling factor; The product of the first decoupling factor and the second decoupling factor is used as the decoupling degree index.

6. The hybrid power system configuration evaluation and parameter optimization method according to claim 1, characterized in that, Under the target configuration, a time-series power sequence is generated based on the system operating data, and a resampling operation is performed on the time-series power sequence to obtain a resampled power sequence, specifically including: Extract the target operating data corresponding to the target configuration from the system operating data, and extract the engine power value and motor power value from the target operating data in time order to generate the time-series power sequence; The time-series power sequence is uniformly resampled along the time axis, with the resampling time interval being a preset time step, to obtain an initial resampled sequence with uniform time intervals. A sliding window filter is applied to the initial resampled sequence, and the median value within each window is taken as the output value of that window. The sliding step size is one time step, thus obtaining the resampled power sequence.

7. The hybrid power system configuration evaluation and parameter optimization method according to claim 1, characterized in that, Based on the resampled power sequence, a set of parameters to be optimized is defined, specifically including: The set of parameters to be optimized includes engine operating point parameters and electric motor operating point parameters; For each power value in the resampled power sequence, the preset engine universal characteristic data is searched to obtain the engine speed range and engine torque range corresponding to the power value. A first speed value is selected from the engine speed range as the engine operating point parameter, and a first torque value is selected from the engine torque range as the motor operating point parameter; Substitute the first speed value into the engine universal characteristic data to query the corresponding engine fuel consumption rate, and substitute the first torque value into the preset motor efficiency characteristic data to query the corresponding motor efficiency value. The engine speed values ​​with fuel consumption rates lower than a preset fuel consumption threshold are retained as candidate engine operating point parameters, and the torque values ​​with motor efficiency values ​​higher than a preset efficiency threshold are retained as candidate motor operating point parameters. The set of parameters to be optimized is composed of all candidate engine operating point parameters and candidate motor operating point parameters.

8. The method for evaluating the configuration and optimizing the parameters of a hybrid power system according to claim 1, characterized in that, Based on the set of parameters to be optimized and the design structure matrix of the target configuration, a parameter evaluation matrix is ​​constructed. Matrix decomposition is performed on the parameter evaluation matrix, and the adjustment direction of each parameter to be optimized is determined using the decomposition results. Specifically, this includes: Each parameter in the set of parameters to be optimized is used as a row index, and each component in the design structure matrix of the target configuration is used as a column index to construct an initial parameter evaluation matrix. The matrix elements are the sensitivity coefficients of the corresponding parameters to the corresponding components. Perform orthogonal triangular decomposition on the initial parameter evaluation matrix to obtain an upper triangular matrix, and extract the diagonal elements of the upper triangular matrix as the independent contribution of each parameter; The independent contributions are sorted in descending order, and the parameters corresponding to the independent contributions with odd-numbered sorting numbers are marked as positive adjustment parameters, while the parameters corresponding to the independent contributions with even-numbered sorting numbers are marked as negative adjustment parameters. The direction of increasing the positive adjustment parameter is taken as the first adjustment direction, and the direction of decreasing the negative adjustment parameter is taken as the second adjustment direction.

9. The hybrid power system configuration evaluation and parameter optimization method according to claim 8, characterized in that, Following the aforementioned adjustment direction, the set of parameters to be optimized is iteratively updated until a preset iteration termination condition is met, resulting in an optimized set of parameters, specifically including: Increase all parameters in the first adjustment direction by a first preset step size, and decrease all parameters in the second adjustment direction by a second preset step size to obtain an updated set of parameters to be optimized; Substitute the updated set of parameters to be optimized into the system simulation model of the target configuration to calculate the overall system efficiency; The overall system efficiency is compared with the overall system efficiency of the previous iteration. If the overall system efficiency is improved, the current update is retained. If the overall system efficiency is reduced, the system reverts to the parameter set of the previous iteration and the first preset step size and the second preset step size are both reduced to half of their original values. When both the first preset step size and the second preset step size are less than the preset step size threshold, the preset iteration termination condition is met, and the current parameter set is used as the optimized parameter set.

10. The method for evaluating the configuration and optimizing the parameters of a hybrid power system according to claim 7, characterized in that, The preset engine universal characteristic data includes braking ratio fuel consumption data under different speed and torque combinations, and the preset motor efficiency characteristic data includes motor system efficiency data under different speed and torque combinations.

Images