Energy consumption prediction method and system for driving path of electric vehicle

Abstract

The invention relates to the technical field of data prediction analysis, in particular to an energy consumption prediction method and system for an electric vehicle driving path, and the method comprises the following steps: collecting a signal through a suspension displacement sensor to generate static whole vehicle mass, monitoring motor torque and longitudinal actual acceleration, and calculating theoretical longitudinal acceleration; and calculating an acceleration residual error, inputting the acceleration residual error into a recursive least square algorithm to generate real-time vehicle mass, analyzing a planned path, extracting a slope angle, calculating resistance by using the real-time vehicle mass and the slope angle, generating mechanical power consumption, and finally converting the mechanical power consumption into an energy consumption predicted value. According to the method, the vehicle load state is dynamically and accurately tracked by identifying the real-time whole vehicle mass, the load fluctuation and the road topographic features are brought into the core calculation logic, it is ensured that the energy consumption prediction model can respond to the vehicle load and road condition changes in real time, accumulated deviation caused by fixed constant calculation is effectively eliminated, and the energy consumption prediction accuracy is improved. And the endurance mileage estimation accuracy in a complex driving environment is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of data prediction and analysis technology, and in particular to a method and system for predicting the energy consumption of an electric vehicle's travel path. Background Technology

[0002] The field of data predictive analytics technology mainly involves the use of computer processing units to collect, store, and perform logical operations on massive amounts of business data. It typically encompasses hardware facilities such as central processing units (CPUs), memory, database servers, and input / output interfaces. Traditional energy consumption prediction methods for electric vehicle routes involve a server obtaining the vehicle's origin and destination coordinates via an onboard GPS module, calculating the planned route mileage between the two points using an electronic map database, retrieving a fixed average energy consumption constant per 100 kilometers for the vehicle model from a vehicle parameter database, performing multiplication operations using the CPU to directly derive the estimated energy consumption, and displaying the specific numerical results on an onboard display terminal.

[0003] Existing technologies rely solely on fixed energy consumption constants and mileage for static estimation, neglecting the decisive impact of actual load variations on driving resistance. They also fail to incorporate changes in gravitational potential energy caused by road gradients into energy consumption considerations. Fixed parameters cannot characterize the dynamic mechanical power consumption characteristics of vehicles under different loads and complex road conditions, resulting in severely distorted predicted values ​​under specific conditions such as fully loaded uphill climbing or long downhill driving. The single calculation dimension is difficult to adapt to the ever-changing actual driving environment, leading to insufficient accuracy in displaying remaining range. This can mislead drivers into making incorrect plans for their journeys, potentially causing the vehicle to run out of power or break down midway. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a method and system for predicting the energy consumption of electric vehicle travel routes.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for predicting the energy consumption of an electric vehicle's travel path, comprising the following steps: S1: The suspension displacement voltage signal is collected by the suspension displacement sensor, converted into a suspension displacement value, and the suspension displacement value is multiplied by the spring stiffness coefficient to generate a wheel load value. The wheel load value is added to the unloaded curb weight to generate the static vehicle weight. S2: Monitor the motor torque and actual longitudinal acceleration during straight-line driving, convert the motor torque into driving force using the wheel rolling radius, and calculate the theoretical longitudinal acceleration based on the static vehicle mass, driving force, and Newton's second law; S3: Calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration, input the driving force, the actual longitudinal acceleration and the acceleration residual into the recursive least squares algorithm, and use the static vehicle mass as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass; S4: Analyze the planned driving path to extract the path slope angle and micro-segment length. Calculate the slope resistance and rolling resistance using the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient. Multiply the sum of the slope resistance and rolling resistance by the micro-segment length to generate mechanical power consumption. Convert the mechanical power consumption into an energy consumption prediction value based on the motor energy conversion efficiency.

[0006] As a further aspect of the present invention, step S1 specifically comprises: S11: Activate the suspension displacement sensor with a preset high-frequency sampling rate to synchronously collect the suspension displacement voltage signals corresponding to the four wheels, call the analog-to-digital conversion module to discretize the collected analog voltage signals into digital signals, and use the calibrated linear conversion coefficients to map the digital signals into the suspension displacement values ​​in physical dimensions. S12: Obtain the spring stiffness coefficient of the vehicle suspension system, perform a multiplication operation between the spring stiffness coefficient and the suspension displacement value, calculate the elastic restoring force generated by the compression or tension of the suspension spring, and use it as the dynamic load component borne by multiple wheels to generate the wheel load value. S13: Sum the wheel load values ​​corresponding to the four wheels to obtain the total vehicle load increment, retrieve the pre-stored unloaded curb weight, and add the total vehicle load increment to the unloaded curb weight to construct the static vehicle weight.

[0007] As a further aspect of the present invention, step S2 specifically comprises: S21: The motor controller reads the current motor torque of the motor output shaft in real time, and performs a division operation on the motor torque using the preset wheel rolling radius to convert the rotational torque into a longitudinal traction force acting on the contact point between the wheel and the ground to generate driving force. S22: The vehicle's actual longitudinal acceleration under the current driving state is captured in real time using an on-board precision inertial measurement unit or longitudinal acceleration sensor. At the same time, based on the static vehicle mass and driving force, an ideal dynamic equation is constructed based on Newton's second law. S23: Divide the driving force by the static vehicle mass, ignore the road slope and air resistance interference, and calculate the theoretical motion response that the vehicle should produce when it is only subjected to the driving force on an ideal flat road to generate the theoretical longitudinal acceleration.

[0008] As a further aspect of the present invention, step S3 specifically comprises: S31: Subtract the actual longitudinal acceleration from the theoretical longitudinal acceleration to obtain the deviation value between the two in order to quantify the inconsistency between the current model parameters and the actual system response and generate the acceleration residual; S32: Construct a recursive least squares algorithm model based on the forgetting factor, set the regression vector of the recursive least squares algorithm model as the longitudinal actual acceleration, the observation value as the driving force, and set the static vehicle mass as the initial estimate of the parameter to be identified. S33: The parameter estimate is corrected and updated using the acceleration residual. The gain matrix and covariance matrix are iteratively calculated at each sampling time to gradually converge the parameter estimate to the true physical value, and the real-time vehicle mass is output.

[0009] As a further aspect of the present invention, step S4 specifically comprises: S41: Obtain geographic information data of the planned driving route from the navigation system, use the path discretization algorithm to divide the entire path into several micro-segments, extract the elevation change information and horizontal distance information of each micro-segment, and use trigonometric function operations to parse the path slope angle and the length of the micro-segment corresponding to multiple segments. S42: Call the real-time vehicle mass, combine it with the path slope angle and the preset gravitational acceleration constant and rolling resistance coefficient, and calculate the slope resistance work done by the vehicle to overcome the gravity component and the rolling resistance work done to overcome the tire deformation to generate slope resistance and rolling resistance. S43: Add the slope resistance and rolling resistance to obtain the total driving resistance, and use the total driving resistance multiplied by the length of the micro-segment to calculate the mechanical energy output required for the vehicle to drive on the current micro-segment and generate mechanical power consumption. S44: Query the efficiency MAP of the motor under different operating conditions to obtain the current energy conversion efficiency of the motor. Divide the mechanical power consumption by the energy conversion efficiency of the motor to reverse estimate the amount of electrical energy consumption that the battery needs to provide. Add up the electrical energy consumption of all micro-segments to generate the predicted energy consumption value.

[0010] As a further aspect of the present invention, the parameter identification in S33 includes: Obtain the parameter estimation vector and the covariance matrix from the previous time step, calculate the gain matrix by combining it with the current regression vector, and use the gain matrix and the acceleration residual to correct the parameter estimation vector from the previous time step to generate the intermediate estimate value for the current time step. The covariance matrix is ​​updated according to a preset forgetting factor, and the intermediate estimate is weighted and smoothed using the updated covariance matrix to remove random fluctuations caused by measurement noise. Determine whether the rate of change of the estimated parameter is less than a preset convergence threshold. If the condition is met, lock the current estimated value as the identification result; otherwise, use the current estimated value as the prior value for the next moment and continue iterating until the convergence condition is met to generate the real-time vehicle quality.

[0011] As a further aspect of the present invention, the energy consumption prediction value generation process in S44 includes: The mechanical power consumption and the energy conversion efficiency of the motor are extracted, and an energy consumption calculation model including the conversion loss from mechanical energy to electrical energy is constructed based on the law of conservation of energy. Its mathematical expression is as follows: ; Based on the energy consumption calculation model, the calculated power consumption of each micro-segment is summed to cover the energy demand of the entire planned path, and the predicted energy consumption value is generated. in, This represents the predicted energy consumption value. Represents the total number of micro-road segments. Representing the Mechanical power consumption of a micro-segment Representing the The energy conversion efficiency of the motor corresponding to each micro-segment.

[0012] As a further aspect of the present invention, the process of generating slope resistance and rolling resistance in S42 includes: Based on the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient, a formula for calculating slope resistance is constructed: ; The formula for calculating rolling resistance: ; Based on the slope resistance calculation formula and the rolling resistance calculation formula, the values ​​are substituted into the calculation to quantify the two types of key physical resistances that the vehicle needs to overcome when driving on the target slope road surface, and generate slope resistance and rolling resistance. in, Represents slope resistance. Represents rolling resistance. This represents the real-time vehicle mass. Represents the gravitational acceleration constant. This represents the slope angle of the path. This represents the rolling resistance coefficient.

[0013] As a further aspect of the present invention, the extraction process of the path slope angle and the micro-segment length in S41 includes: The three-dimensional coordinate point set of the planned driving path is analyzed, the altitude difference and horizontal projection distance between adjacent coordinate points are calculated, and the arctangent of the ratio of altitude difference to horizontal projection distance is calculated using inverse trigonometric functions to generate the slope angle of the path. The straight-line distance between adjacent coordinate points in three-dimensional space is calculated using the Euclidean distance formula, or the actual driving mileage is calculated along the path curve based on the integral algorithm. The distance value is used as the differential unit of energy integration to generate the length of the micro-segment. The generated path slope angle sequence is smoothed and filtered to remove high-frequency jitter caused by noise in the map data, while retaining low-frequency trend terms that reflect the actual terrain undulations.

[0014] An energy consumption prediction system for electric vehicle travel routes, the system being used to implement the aforementioned energy consumption prediction method for electric vehicle travel routes, the system comprising: The static mass calculation module is used to collect suspension displacement voltage signals through suspension displacement sensors, convert them into suspension displacement values, multiply the suspension displacement values ​​by the spring stiffness coefficient to generate wheel load values, and accumulate the wheel load values ​​and the unloaded curb weight to generate the static vehicle mass. The theoretical acceleration calculation module is used to monitor the motor torque and actual longitudinal acceleration during straight-line driving. It converts the motor torque into driving force using the wheel rolling radius and calculates the theoretical longitudinal acceleration based on the static vehicle mass, driving force, and Newton's second law. The real-time mass identification module is used to calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration. The driving force, the actual longitudinal acceleration and the acceleration residual are input into the recursive least squares algorithm, and the static vehicle mass is used as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass. The energy consumption prediction generation module is used to analyze the planned driving path to extract the path slope angle and micro-segment length. It calculates the slope resistance and rolling resistance using the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient. The sum of the slope resistance and rolling resistance is multiplied by the micro-segment length to generate mechanical power consumption. Based on the motor energy conversion efficiency, the mechanical power consumption is converted into an energy consumption prediction value.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the static vehicle mass is established by collecting suspension displacement signals, and the real-time vehicle mass is identified by using a recursive least squares algorithm in combination with motor torque and longitudinal acceleration, thereby achieving dynamic and accurate tracking of vehicle load status. The downhill resistance and rolling resistance of micro-segments are calculated by combining the path slope angle, and mechanical power consumption is generated based on the principle of physical work and converted into energy consumption prediction value. Load fluctuations and road terrain features are incorporated into the core calculation logic to ensure that the energy consumption prediction model can respond to changes in vehicle load and road conditions in real time, effectively eliminating the cumulative deviation caused by fixed constant calculation, significantly improving the accuracy of range estimation in complex driving environments, and providing drivers with a reliable basis for trip planning. Attached Figure Description

[0016] Figure 1 This is a flowchart of the energy consumption prediction method for the driving path of an electric vehicle according to the present invention. Figure 2 This is a flowchart illustrating the static vehicle weight generation process of this invention. Figure 3 This is a flowchart illustrating the theoretical longitudinal acceleration calculation of this invention. Figure 4 This is a flowchart of the real-time vehicle quality identification process of the present invention; Figure 5 This is a flowchart illustrating the energy consumption prediction value generation process of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the software-based technical solution is described in detail below with reference to system architecture diagrams and embodiments. It should be understood that the specific embodiments described herein are only for explaining the technical solutions of this invention and do not constitute a limitation on the scope of protection.

[0018] In the description of this invention, the system architecture relationships or data processing flows indicated by terms such as "layer," "module," "interface," "data flow," "client," and "server" are all defined based on the architecture diagram or flowchart corresponding to the embodiments. This way of describing is only used to clearly illustrate the logical relationships between the elements in the technical solution, and not to limit the physical deployment form. The term "multiple" includes two or more technical units, including but not limited to multiple data nodes, processing threads, service instances, or functional components and other scalable elements. The specific number is determined according to the actual business scenario and needs to be specifically specified.

[0019] Please see Figure 1 and Figure 2 This invention provides a technical solution: a method for predicting the energy consumption of an electric vehicle's travel path, comprising the following steps: S1: The suspension displacement voltage signal is collected by the suspension displacement sensor and converted into a suspension displacement value. The suspension displacement value is multiplied by the spring stiffness coefficient to generate the wheel load value. The wheel load value is added together with the unloaded curb weight to generate the static vehicle weight. The specific steps of S1 are as follows: S11: Activate the suspension displacement sensor with a preset high-frequency sampling rate to synchronously collect the suspension displacement voltage signals corresponding to the four wheels, call the analog-to-digital conversion module to discretize the collected analog voltage signals into digital signals, and use the calibrated linear conversion coefficients to map the digital signals into suspension displacement values ​​in physical dimensions. S12: Obtain the spring stiffness coefficient of the vehicle suspension system, perform a multiplication operation between the spring stiffness coefficient and the suspension displacement value, calculate the elastic restoring force generated by the compression or tension of the suspension spring, and use it as the dynamic load component borne by multiple wheels to generate wheel load values. S13: Sum the wheel load values ​​corresponding to the four wheels to obtain the total vehicle load increment, retrieve the pre-stored unloaded curb weight, and add the total vehicle load increment to the unloaded curb weight to construct the static vehicle weight.

[0020] In this embodiment, a four-wheel drive electric vehicle was selected as the test subject. The vehicle's factory-specified unloaded curb weight was... The load is 1900.0 kg. To achieve accurate sensing of the vehicle's load, linear resistive suspension displacement sensors were installed at the four suspension systems: left front, right front, left rear, and right rear. The power supply voltage for the sensors is... The measured stroke is to The linear output voltage range is to .

[0021] In step S11, the vehicle is in a stationary power-on state, and the onboard central processing unit... The high-frequency sampling rate activates the analog-to-digital converter (ADC). The ADC uses a 12-bit precision conversion chip to synchronously discretize the analog voltage signals from the four suspension displacement sensors into digital signals. The processor then calls preset calibration parameters, including linear conversion coefficients. and zero intercept (This parameter is calibrated by measuring the voltage value when the vehicle is placed on a lift and the suspension is fully suspended.) At the current sampling moment, the collected and filtered voltage signals of the four wheels are as follows: Left front , right front Left rear Right rear Calculate the suspension displacement values ​​in terms of physical dimensions using a linear mapping relationship: Left front suspension displacement ; Right front suspension displacement ; Left rear suspension displacement ; Right rear suspension displacement .

[0022] In step S12, the system retrieves the spring stiffness coefficients of the vehicle suspension system from the read-only memory. Considering the vehicle's load distribution characteristics, the front suspension spring stiffness coefficients... Set as Rear suspension spring stiffness coefficient Set as The processor performs multiplication to calculate the elastic restoring force of each suspension spring due to its current compression, i.e., the wheel load value: Left front wheel load ; Right front wheel load ; Left rear wheel load ; Right rear wheel load .

[0023] In step S13, the load values ​​of the four wheels are first summed to obtain the total support force of the entire vehicle. Subsequently, using the gravitational acceleration constant... Convert the total support force into the current total physical mass, including the unloaded mass. It should be noted that due to the different zero-point calibration positions of the sensors (calibrated in this example with the suspension fully extended), the calculated values ​​will differ. Correction needs to be performed using baseline compression data under no-load conditions, or an incremental method can be used directly. In this embodiment, the incremental method logic is adopted: the system records the baseline total load under no-load conditions as 12740N (corresponding to 1300kg of sprung mass, with the remaining 600kg being unsprung mass not supported by the suspension). The current total load increment is... Calculate the incremental mass of the load. Finally, this increment is added to the unloaded curb weight to construct the static vehicle weight. This figure accurately reflects the vehicle's actual stationary mass when carrying approximately four adult passengers and a small amount of luggage.

[0024] The aforementioned suspension displacement sensor refers to a linear potentiometer or Hall effect displacement sensor installed in the automotive suspension system to measure the change in the vertical distance between the wheel and the vehicle body in real time. Its output signal is usually an analog voltage or a PWM waveform.

[0025] Please see Figure 1 and Figure 3 S2: Monitor the motor torque and actual longitudinal acceleration during straight-line driving, convert the motor torque into driving force using the wheel rolling radius, and calculate the theoretical longitudinal acceleration based on the static vehicle mass, driving force and Newton's second law; The specific steps of S2 are as follows: S21: The motor controller reads the current motor torque of the motor output shaft in real time, and performs a division operation on the motor torque using the preset wheel rolling radius to convert the rotational torque into a longitudinal traction force acting on the contact point between the wheel and the ground to generate driving force. S22: Real-time capture of the vehicle's actual longitudinal acceleration under the current driving state using an on-board precision inertial measurement unit or longitudinal acceleration sensor, and construction of ideal dynamic equations based on Newton's second law based on the static vehicle mass and driving force. S23: Divide the driving force by the static vehicle mass, ignore the road slope and air resistance interference, calculate the theoretical motion response that the vehicle should produce when it is only subjected to the driving force on an ideal flat road, and generate the theoretical longitudinal acceleration.

[0026] In step S21, the vehicle enters the straight-line acceleration section, and the vehicle's computer monitors the operating data broadcast by the motor controller via the CAN bus in real time. When it is determined that the vehicle's steering wheel angle is less than... Furthermore, when the accelerator pedal opening is stable, the system locks the current moment. motor torque Assume the collected motor torque is... The system simultaneously retrieves powertrain parameters: main reducer speed ratio. Mechanical transmission efficiency and the calibrated wheel rolling radius The processor uses division and multiplication operations to convert the rotational torque into a longitudinal driving force acting on the ground. : .

[0027] This driving force value represents the theoretical total thrust applied to the vehicle by the powertrain.

[0028] In step S22, the vehicle's current longitudinal acceleration is captured in real time using an onboard high-precision six-axis inertial measurement unit (IMU). To ensure data integrity, a low-pass filter is integrated into the IMU, with the cutoff frequency set to [value missing]. This is to filter out high-frequency noise caused by road bumps. At any given time... The actual longitudinal acceleration after filtering Measured as Simultaneously, the system retrieves the static vehicle mass calculated in step S1. As the initial mass parameter of the dynamic equation.

[0029] In step S23, according to Newton's second law We construct the motion equations under an ideal, straight road surface, neglecting the instantaneous effects of wind resistance and rolling resistance, and calculate the vehicle's motion under conditions only affected by wind resistance and rolling resistance. The theoretical longitudinal acceleration that should be generated under action The calculation process is as follows: .

[0030] At this moment, the system obtained two acceleration values ​​at the same instant: one calculated based on the static model. Compared with the actual measurement of the sensor The difference between the two indicates that the static weighing mass cannot fully represent the equivalent inertial characteristics of a vehicle after overcoming air resistance, gradient resistance, and acceleration inertial resistance during dynamic driving, and parameter identification is required in the next step.

[0031] Table 1 shows the results of key parameter collection and calculation for stages S1 and S2.

[0032] Table 1 Calculation of Vehicle Mass and Dynamic Parameters

[0033] The aforementioned inertial measurement unit refers to a microelectromechanical system (MEMS) sensor combination, which typically includes a three-axis accelerometer and a three-axis gyroscope. It can measure the specific force (acceleration) and angular velocity of an object in three-dimensional space, and calculate the object's velocity, position, and attitude angle through integration.

[0034] Please see Figure 1 and Figure 4 S3: Calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration. Input the driving force, actual longitudinal acceleration and acceleration residual into the recursive least squares algorithm. Use the static vehicle mass as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass. The specific steps for S3 are as follows: S31: Subtract the actual longitudinal acceleration from the theoretical longitudinal acceleration to obtain the deviation between the two in order to quantify the inconsistency between the current model parameters and the actual system response and generate acceleration residuals; S32: Construct a recursive least squares algorithm model based on the forgetting factor, set the regression vector of the recursive least squares algorithm model as the longitudinal actual acceleration, the observation value as the driving force, and set the static vehicle mass as the initial estimate of the parameter to be identified. S33: The parameter estimates are corrected and updated using acceleration residuals. The gain matrix and covariance matrix are iteratively calculated at each sampling time, so that the parameter estimates gradually converge to the true physical values ​​and the real-time vehicle mass is output. S33 parameter identification includes: Obtain the parameter estimation vector and covariance matrix from the previous time step, calculate the gain matrix by combining it with the current regression vector, and use the gain matrix and acceleration residual to correct the parameter estimation vector from the previous time step to generate the intermediate estimate value for the current time step. The covariance matrix is ​​updated based on a preset forgetting factor, and the intermediate estimates are weighted and smoothed using the updated covariance matrix to remove random fluctuations caused by measurement noise. Determine whether the rate of change of the parameter estimate is less than the preset convergence threshold. If the condition is met, lock the current estimate as the identification result. Otherwise, use the current estimate as the prior value for the next moment and continue iterating until the convergence condition is met to generate the real-time vehicle quality.

[0035] In step S31, the processor first performs a subtraction operation to calculate the deviation between the theoretical longitudinal acceleration and the actual longitudinal acceleration, i.e., the acceleration residual. Based on the aforementioned data: .

[0036] The positive residual indicates that the actual acceleration experienced by the vehicle under the current driving force is less than the expected value based on static mass, suggesting the existence of unmodeled drag or an increase in equivalent mass within the system.

[0037] In step S32, the system initializes the recursive least squares (FF-RLS) algorithm model based on the forgetting factor. The regression vector is set. The actual longitudinal acceleration Observed values As driving force Parameters to be identified Defined as "real-time vehicle weight" Its physical meaning includes the vehicle's physical mass and the generalized mass after converting driving resistance into inertial load. The static vehicle mass obtained in step S1... Set as the initial estimate of the parameter to be identified Simultaneously, initialize the covariance matrix. Forgetting factor It was set to 0.98 to balance the algorithm's ability to remember historical data with its ability to track new changes.

[0038] In step S33, the acceleration residual is used to drive the correction update of the parameter estimate.

[0039] First, calculate the gain matrix by combining the current regression vector with the covariance matrix from the previous time step. : ; Substitute the values: .

[0040] Secondly, the parameter estimates are corrected using the gain matrix. The correction term is the gain matrix multiplied by the prior prediction error (here, the prior prediction error is used). and (Approximate representation of the difference) Prediction error .

[0041] Update parameter estimates: .

[0042] Subsequently, the covariance matrix is ​​updated. For use in the next iteration: .

[0043] The sharp decrease in the value of the covariance matrix indicates a significant improvement in the algorithm's confidence in the parameter estimates. Finally, the system calculates the rate of change of the parameter estimates. Since the rate of change exceeded the preset convergence threshold of 1%, the system used 2591.9 kg as the prior value for the next time step and continued iterating. After approximately 10 consecutive sampling cycles of iteration, the parameter estimate finally converged and stabilized at 2650.0 kg. The system locked this value as the real-time vehicle weight. Output.

[0044] The recursive least squares algorithm described above is an adaptive filter algorithm. It recursively updates the parameter estimates based on new data samples by minimizing the weighted least squares cost function. It does not require storing and processing all historical data, making it very suitable for online real-time parameter identification.

[0045] Please see Figure 1 and Figure 5 S4: Analyze the planned driving path to extract the path slope angle and micro-segment length. Calculate the slope resistance and rolling resistance using real-time vehicle mass, path slope angle, gravitational acceleration constant, and rolling resistance coefficient. Multiply the sum of slope resistance and rolling resistance by the micro-segment length to generate mechanical power consumption. Convert the mechanical power consumption into an energy consumption prediction value based on the motor energy conversion efficiency.

[0046] The specific steps for S4 are as follows: S41: Obtain geographic information data of the planned driving route from the navigation system, use the path discretization algorithm to divide the entire route into several micro-segments, extract the elevation change information and horizontal distance information of each micro-segment, and use trigonometric function operations to parse out the path slope angle and micro-segment length corresponding to multiple segments. The process of extracting the path slope angle and micro-segment length of S41 includes: The three-dimensional coordinate point set of the planned driving path is analyzed, the altitude difference and horizontal projection distance between adjacent coordinate points are calculated, and the arctangent of the ratio of altitude difference to horizontal projection distance is calculated using inverse trigonometric functions to generate the path slope angle. The straight-line distance between adjacent coordinate points in three-dimensional space is calculated using the Euclidean distance formula, or the actual driving mileage is calculated along the path curve based on the integral algorithm. The distance value is used as the differential unit of energy integration to generate the micro-segment length. The generated path slope angle sequence is smoothed and filtered to remove high-frequency jitter caused by noise in the map data and retain low-frequency trend terms that reflect the real terrain undulations. S42: Call the real-time vehicle mass, combine the path slope angle and the preset gravitational acceleration constant and rolling resistance coefficient, calculate the slope resistance work done by the vehicle to overcome the gravity component and the rolling resistance work done to overcome the tire deformation to generate slope resistance and rolling resistance. The generation process of slope resistance and rolling resistance of S42 includes: By obtaining real-time vehicle mass, path slope angle, gravitational acceleration constant, and rolling resistance coefficient, a formula for calculating slope resistance is constructed: ; The formula for calculating rolling resistance: ; By substituting the numerical values ​​into the formulas for calculating slope resistance and rolling resistance respectively, the two types of key physical resistances that a vehicle needs to overcome when driving on a target slope road surface are quantified, and slope resistance and rolling resistance are generated. in, Represents slope resistance. Represents rolling resistance. Represents real-time vehicle weight. Represents the gravitational acceleration constant. Represents the slope angle of the path. Represents the rolling resistance coefficient; S43: Add the gradient resistance and rolling resistance to obtain the total driving resistance. Multiply the total driving resistance by the length of the micro-segment to calculate the mechanical energy output required for the vehicle to drive on the current micro-segment and generate mechanical power consumption. S44: Query the efficiency MAP of the motor under different operating conditions to obtain the current motor energy conversion efficiency. Divide the mechanical power consumption by the motor energy conversion efficiency to reverse calculate the amount of electrical energy consumption that the battery needs to provide. Then, add up the electrical energy consumption of all micro-segments to generate the energy consumption prediction value. The energy consumption prediction generation process for S44 includes: By extracting mechanical power consumption and motor energy conversion efficiency, and based on the law of conservation of energy, an energy consumption calculation model is constructed that includes the loss in the conversion of mechanical energy to electrical energy. Its mathematical expression is as follows: ; Based on the energy consumption calculation model, the calculated power consumption of each micro-segment is summed to cover the energy demand of the entire planned route and generate energy consumption prediction values. in, Represents the predicted energy consumption value. Represents the total number of micro-road segments. Representing the Mechanical power consumption of a micro-segment Representing the The energy conversion efficiency of the motor corresponding to each micro-segment.

[0047] In step S41, the vehicle navigation system plans a 2km long test route. The route discretization algorithm divides this route into 20 micro-segments. Taking the first micro-segment as an example, the system extracts the 3D coordinate data of the start and end points from a high-precision map database. (Starting point elevation...) Ending altitude Horizontal projection distance .

[0048] Using inverse trigonometric functions to analyze the slope angle of the path : (about ).

[0049] Calculating the length of a micro-segment using the Euclidean distance formula : .

[0050] To eliminate noise in the map data, the system performed a five-point moving average filter on the extracted slope sequence.

[0051] In step S42, the real-time vehicle mass, which was stably output in step S3, is retrieved. Simultaneously, the gravitational acceleration constant is obtained. and the rolling resistance coefficient under current road conditions (Corresponding to dry asphalt pavement). The processor calculates the gradient resistance that the vehicle needs to overcome in this micro-road section. and rolling resistance .

[0052] Calculate slope resistance: .

[0053] Calculate rolling resistance: .

[0054] This result indicates that, in the current On uphill sections, gradient resistance dominates the driving resistance. If calculated using an uncorrected static mass (2175kg), the gradient resistance is only 852N, which would lead to a significant negative bias in energy consumption prediction.

[0055] In step S43, the two types of resistance are added together to obtain the total driving resistance. The mechanical power consumption is calculated by multiplying the total resistance by the length of the micro-segment. : .

[0056] Convert it to kilowatt-hours: .

[0057] In step S44, the system determines the required torque based on the current driving conditions (vehicle speed 60km / h, corresponding torque). Query the efficiency MAP of the motor. At this operating point, the motor's energy conversion efficiency... It was calibrated to 0.90. The required electrical energy output from the battery is calculated by using mechanical power consumption. : .

[0058] Finally, the system sums up the energy consumption of all 20 micro-segments. Assuming the total energy consumption of the remaining 19 micro-segments is 0.8500 kWh, the predicted energy consumption for the entire process is... for: .

[0059] Experimental data show that, compared with the traditional method that only uses static mass for prediction (predicted value is 0.73kWh), the relative error of the energy consumption prediction value based on real-time vehicle mass identification proposed in this invention is reduced from 18% to less than 3%, which greatly improves the accuracy of the remaining driving range display.

[0060] Table 2 shows the detailed energy consumption prediction data for a certain segment of the path in this embodiment.

[0061] Table 2. Energy Consumption Prediction Results for Micro-Segments of the Route

[0062] The rolling resistance coefficient mentioned above refers to the ratio of the resistance generated by the tire's internal friction, hysteresis loss, and road deformation to the normal load on the wheel when the wheel rolls on the road surface. Its value depends on the tire pressure, tread pattern, and road surface roughness.

[0063] in, Represents slope resistance. Represents rolling resistance. Represents real-time vehicle weight. Represents the gravitational acceleration constant. Represents the slope angle of the path. Represents the rolling resistance coefficient. Represents the predicted energy consumption value. Representing the Mechanical power consumption of a micro-segment Representing the The energy conversion efficiency of the motor corresponding to each micro-segment.

[0064] An energy consumption prediction system for electric vehicle travel routes, the system being used to execute the aforementioned energy consumption prediction method for electric vehicle travel routes, the system comprising: The static mass calculation module is used to collect suspension displacement voltage signals through suspension displacement sensors, convert them into suspension displacement values, multiply the suspension displacement values ​​by the spring stiffness coefficient to generate wheel load values, and accumulate the wheel load values ​​and the unloaded curb weight to generate the static vehicle mass. The theoretical acceleration calculation module is used to monitor the motor torque and actual longitudinal acceleration during straight-line driving. It converts the motor torque into driving force using the wheel rolling radius and calculates the theoretical longitudinal acceleration based on the static vehicle mass, driving force, and Newton's second law. The real-time mass identification module is used to calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration. The driving force, actual longitudinal acceleration and acceleration residual are input into the recursive least squares algorithm. The static vehicle mass is used as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass. The energy consumption prediction generation module is used to analyze the planned driving path to extract the path slope angle and micro-segment length. It calculates the slope resistance and rolling resistance using real-time vehicle mass, path slope angle, gravitational acceleration constant and rolling resistance coefficient. The sum of slope resistance and rolling resistance is multiplied by the micro-segment length to generate mechanical power consumption. Based on the motor energy conversion efficiency, the mechanical power consumption is converted into an energy consumption prediction value.

[0065] The above embodiments illustrate preferred embodiments of the present invention. Any equivalent adjustments to the technical solution based on software engineering methods are within the scope of protection, including but not limited to: implementing algorithm logic using different programming languages, refactoring functional modules into services, adjusting data interaction protocols, and optimizing resource scheduling strategies. Any implementation scheme derived from reasonable modifications to the data processing flow, service call chain, or system architecture layer without departing from the core technology of the present invention should be considered within the protection scope defined by the technical solution of the present invention.

Claims

1. A method for predicting energy consumption along the travel path of an electric vehicle, characterized in that, Includes the following steps: S1: The suspension displacement voltage signal is collected by the suspension displacement sensor, converted into a suspension displacement value, and the suspension displacement value is multiplied by the spring stiffness coefficient to generate a wheel load value. The wheel load value is added to the unloaded curb weight to generate the static vehicle weight. S2: Monitor the motor torque and actual longitudinal acceleration during straight-line driving, convert the motor torque into driving force using the wheel rolling radius, and calculate the theoretical longitudinal acceleration based on the static vehicle mass, driving force, and Newton's second law; S3: Calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration, input the driving force, the actual longitudinal acceleration and the acceleration residual into the recursive least squares algorithm, and use the static vehicle mass as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass; S4: Analyze the planned driving path to extract the path slope angle and micro-segment length. Calculate the slope resistance and rolling resistance using the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient. Multiply the sum of the slope resistance and rolling resistance by the micro-segment length to generate mechanical power consumption. Convert the mechanical power consumption into an energy consumption prediction value based on the motor energy conversion efficiency.

2. The energy consumption prediction method for electric vehicle travel routes according to claim 1, characterized in that, The specific steps of S1 are as follows: S11: Activate the suspension displacement sensor with a preset high-frequency sampling rate to synchronously collect the suspension displacement voltage signals corresponding to the four wheels, call the analog-to-digital conversion module to discretize the collected analog voltage signals into digital signals, and use the calibrated linear conversion coefficients to map the digital signals into the suspension displacement values ​​in physical dimensions. S12: Obtain the spring stiffness coefficient of the vehicle suspension system, perform a multiplication operation between the spring stiffness coefficient and the suspension displacement value, calculate the elastic restoring force generated by the compression or tension of the suspension spring, and use it as the dynamic load component borne by multiple wheels to generate the wheel load value. S13: Sum the wheel load values ​​corresponding to the four wheels to obtain the total vehicle load increment, retrieve the pre-stored unloaded curb weight, and add the total vehicle load increment to the unloaded curb weight to construct the static vehicle weight.

3. The energy consumption prediction method for the electric vehicle's travel path according to claim 2, characterized in that, The specific steps of S2 are as follows: S21: The motor controller reads the current motor torque of the motor output shaft in real time, and performs a division operation on the motor torque using the preset wheel rolling radius to convert the rotational torque into a longitudinal traction force acting on the contact point between the wheel and the ground to generate driving force. S22: The vehicle's actual longitudinal acceleration under the current driving state is captured in real time using an on-board precision inertial measurement unit or longitudinal acceleration sensor. At the same time, based on the static vehicle mass and driving force, an ideal dynamic equation is constructed based on Newton's second law. S23: Divide the driving force by the static vehicle mass, ignore the road slope and air resistance interference, and calculate the theoretical motion response that the vehicle should produce when it is only subjected to the driving force on an ideal flat road to generate the theoretical longitudinal acceleration.

4. The energy consumption prediction method for the electric vehicle's travel path according to claim 3, characterized in that, The specific steps of S3 are as follows: S31: Subtract the actual longitudinal acceleration from the theoretical longitudinal acceleration to obtain the deviation value between the two in order to quantify the inconsistency between the current model parameters and the actual system response and generate the acceleration residual; S32: Construct a recursive least squares algorithm model based on the forgetting factor, set the regression vector of the recursive least squares algorithm model as the longitudinal actual acceleration, the observation value as the driving force, and set the static vehicle mass as the initial estimate of the parameter to be identified. S33: The parameter estimate is corrected and updated using the acceleration residual. The gain matrix and covariance matrix are iteratively calculated at each sampling time to gradually converge the parameter estimate to the true physical value, and the real-time vehicle mass is output.

5. The energy consumption prediction method for the electric vehicle's travel path according to claim 4, characterized in that, The specific steps of S4 are as follows: S41: Obtain geographic information data of the planned driving route from the navigation system, use the path discretization algorithm to divide the entire path into several micro-segments, extract the elevation change information and horizontal distance information of each micro-segment, and use trigonometric function operations to parse the path slope angle and the length of the micro-segment corresponding to multiple segments. S42: Call the real-time vehicle mass, combine it with the path slope angle and the preset gravitational acceleration constant and rolling resistance coefficient, and calculate the slope resistance work done by the vehicle to overcome the gravity component and the rolling resistance work done to overcome the tire deformation to generate slope resistance and rolling resistance. S43: Add the slope resistance and rolling resistance to obtain the total driving resistance, and use the total driving resistance multiplied by the length of the micro-segment to calculate the mechanical energy output required for the vehicle to drive on the current micro-segment and generate mechanical power consumption. S44: Query the efficiency MAP of the motor under different operating conditions to obtain the current energy conversion efficiency of the motor. Divide the mechanical power consumption by the energy conversion efficiency of the motor to reverse estimate the amount of electrical energy consumption that the battery needs to provide. Add up the electrical energy consumption of all micro-segments to generate the predicted energy consumption value.

6. The energy consumption prediction method for the electric vehicle's travel path according to claim 4, characterized in that, The parameter identification of S33 includes: Obtain the parameter estimation vector and the covariance matrix from the previous time step, calculate the gain matrix by combining it with the current regression vector, and use the gain matrix and the acceleration residual to correct the parameter estimation vector from the previous time step to generate the intermediate estimate value for the current time step. The covariance matrix is ​​updated according to a preset forgetting factor, and the intermediate estimate is weighted and smoothed using the updated covariance matrix to remove random fluctuations caused by measurement noise. Determine whether the rate of change of the estimated parameter is less than a preset convergence threshold. If the condition is met, lock the current estimated value as the identification result; otherwise, use the current estimated value as the prior value for the next moment and continue iterating until the convergence condition is met to generate the real-time vehicle quality.

7. The energy consumption prediction method for the driving path of an electric vehicle according to claim 5, characterized in that, The energy consumption prediction generation process of S44 includes: The mechanical power consumption and the energy conversion efficiency of the motor are extracted, and an energy consumption calculation model including the conversion loss from mechanical energy to electrical energy is constructed based on the law of conservation of energy. Its mathematical expression is as follows: ； Based on the energy consumption calculation model, the calculated power consumption of each micro-segment is summed to cover the energy demand of the entire planned path, and the predicted energy consumption value is generated. in, This represents the predicted energy consumption value. Represents the total number of micro-road segments. Representing the Mechanical power consumption of a micro-segment Representing the The energy conversion efficiency of the motor corresponding to each micro-segment.

8. The energy consumption prediction method for the driving path of an electric vehicle according to claim 5, characterized in that, The process of generating slope resistance and rolling resistance in S42 includes: Based on the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient, a formula for calculating slope resistance is constructed: ； The formula for calculating rolling resistance: ； Based on the slope resistance calculation formula and the rolling resistance calculation formula, the values ​​are substituted into the calculation to quantify the two types of key physical resistances that the vehicle needs to overcome when driving on the target slope road surface, and generate slope resistance and rolling resistance. in, Represents slope resistance. Represents rolling resistance. This represents the real-time vehicle mass. Represents the gravitational acceleration constant. This represents the slope angle of the path. This represents the rolling resistance coefficient.

9. The energy consumption prediction method for electric vehicle travel routes according to claim 5, characterized in that, The extraction process of the path slope angle and the micro-segment length in S41 includes: The three-dimensional coordinate point set of the planned driving path is analyzed, the altitude difference and horizontal projection distance between adjacent coordinate points are calculated, and the arctangent of the ratio of altitude difference to horizontal projection distance is calculated using inverse trigonometric functions to generate the slope angle of the path. The straight-line distance between adjacent coordinate points in three-dimensional space is calculated using the Euclidean distance formula, or the actual driving mileage is calculated along the path curve based on the integral algorithm. The distance value is used as the differential unit of energy integration to generate the length of the micro-segment. The generated path slope angle sequence is smoothed and filtered to remove high-frequency jitter caused by noise in the map data, while retaining low-frequency trend terms that reflect the actual terrain undulations.

10. An energy consumption prediction system for the driving path of an electric vehicle, characterized in that, The system is used to implement the energy consumption prediction method for the electric vehicle driving path according to any one of claims 1-9, and the system includes: The static mass calculation module is used to collect suspension displacement voltage signals through suspension displacement sensors, convert them into suspension displacement values, multiply the suspension displacement values ​​by the spring stiffness coefficient to generate wheel load values, and accumulate the wheel load values ​​and the unloaded curb weight to generate the static vehicle mass. The theoretical acceleration calculation module is used to monitor the motor torque and actual longitudinal acceleration during straight-line driving. It converts the motor torque into driving force using the wheel rolling radius and calculates the theoretical longitudinal acceleration based on the static vehicle mass, driving force, and Newton's second law. The real-time mass identification module is used to calculate the acceleration residual between the theoretical longitudinal acceleration and the actual longitudinal acceleration. The driving force, the actual longitudinal acceleration and the acceleration residual are input into the recursive least squares algorithm, and the static vehicle mass is used as the initial value for iteration to perform parameter identification and generate the real-time vehicle mass. The energy consumption prediction generation module is used to analyze the planned driving path to extract the path slope angle and micro-segment length. It calculates the slope resistance and rolling resistance using the real-time vehicle mass, the path slope angle, the gravitational acceleration constant, and the rolling resistance coefficient. The sum of the slope resistance and rolling resistance is multiplied by the micro-segment length to generate mechanical power consumption. Based on the motor energy conversion efficiency, the mechanical power consumption is converted into an energy consumption prediction value.

Images