Staged power matching control method for sodium ion battery excavator

Abstract

The invention relates to the technical field of engineering machinery control, and discloses a sodium ion battery excavator staged power matching control method which comprises the following steps: acquiring a pressure signal of a main pump of a hydraulic system in real time; the pressure signals are input into a pre-trained hidden Markov model, and the current operation stage of the excavator is recognized in real time; acquiring the real-time load size in the current working stage, comparing the real-time load size with a preset standard load range, and judging the current working condition grade; according to the staged power matching control method for the sodium ion battery excavator, the energy efficiency of the battery can be improved.

Description

Technical Field

[0001] This invention relates to the field of engineering machinery control technology, and in particular to a phased power matching control method for sodium-ion battery excavators. Background Technology

[0002] Based on a systematic analysis of current excavator power matching methods, and combined with the characteristics of sodium-ion batteries and cutting-edge advancements in intelligent state recognition technology, a novel staged power matching control framework has been constructed. Its technical background can be explained from the following four aspects:

[0003] The inherent drawbacks of traditional excavator power matching control strategies: Most mainstream hydraulic excavators currently use load-sensitive control or positive / negative flow control systems. The core principle is to adjust the output flow of the hydraulic pump through a pressure compensation valve to match the actuator's requirements.

[0004] In existing technologies, the power matching control of excavators during operation suffers from the following significant drawbacks: Insufficient dynamic adaptability: Traditional control strategies struggle to accurately respond to load fluctuations at different stages of the work cycle (such as digging, hoisting and slewing, unloading, empty bucket return, and digging preparation). Underutilization of battery performance: While sodium-ion batteries possess high-rate discharge and wide-temperature adaptability, existing control methods lack optimization strategies tailored to their characteristics, failing to dynamically allocate power according to real-time load demands, resulting in wasted battery capacity or overload risks. Imbalance between energy consumption and efficiency: Under conditions of drastic load changes (such as heavy-load digging or light-load slewing), traditional methods rely on fixed power output, leading to increased hydraulic system losses and prolonged overall machine operation cycle time. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a staged power matching control method for sodium-ion battery excavators, which can improve battery energy efficiency.

[0006] This invention provides a staged power matching control method for sodium-ion battery excavators, comprising the following steps: Real-time acquisition of pressure signals from the main pump of the hydraulic system; The pressure signal is input into the hidden Markov model to identify the current operation stage of the excavator in real time. The operation stages include digging, hoisting and slewing, unloading, empty bucket return and digging preparation. Obtain the real-time load size under the current operation stage and compare it with the preset standard load range to determine the current working condition level; Real-time acquisition of the state of charge (SOC) value of sodium-ion batteries; Based on the current operating stage, operating condition level, and state of charge value, the output power of the sodium-ion battery is obtained through a preset dynamic power mapping table.

[0007] Furthermore, the pre-trained Hidden Markov Model is constructed and trained in the following manner: Collect historical main pump pressure data of the excavator in multiple complete work cycles, and clean and normalize the data. The processed historical data was manually annotated, and five hidden states were defined to correspond to five operation stages. The pressure data were converted into a sequence of observation symbols using an equal-width discretization method; The Baum-Welch algorithm is used to iteratively train the probability vector π, state transition matrix A, and observation probability matrix B in the initial hidden Markov model. The trained model was evaluated using five-fold cross-validation.

[0008] Furthermore, it also includes preprocessing the real-time acquired pressure signals. The preprocessing process includes the following steps: Perform low-pass digital filtering on the pressure signal; Based on the pressure signal within the sliding window, its time-domain characteristics are obtained, including the pressure mean, pressure variance, and instantaneous rate of change of pressure. The obtained temporal features are used as the observation input for the Hidden Markov Model.

[0009] Furthermore, the digital filtering employs a Butterworth low-pass filter with a cutoff frequency of 20Hz, and the system delay time is controlled within 50ms.

[0010] Furthermore, the sliding window has a window duration of 50 milliseconds, an overlap rate of 50%, and the temporal features are updated every 50 milliseconds.

[0011] Furthermore, determining the current operating condition level includes: If the real-time pressure is lower than 80% of the lower limit of the standard load range, it is determined to be a light load condition. If the real-time pressure is between 80% of the lower limit and 120% of the upper limit of the standard load range, it is considered to be in normal operating condition. If the real-time pressure is higher than 120% of the upper limit of the standard load range, it is judged as a heavy load condition. The current operating condition is only confirmed when the operating condition level is consistent for three consecutive sampling periods.

[0012] Furthermore, the process of obtaining the state of charge value includes the following steps: Establish a communication connection with the battery management system via CAN bus; Send SOC requests and receive response data at a frequency of not less than 10Hz; Temperature compensation and internal resistance compensation are performed on the obtained SOC value; If communication fails, the historical SOC value will be used as a substitute.

[0013] Furthermore, it also includes a SOC trend prediction step: using a Kalman filter algorithm to predict the SOC change trend in the next 10 minutes, and if the prediction shows that the SOC decline rate exceeds the threshold, then the power back-off protection mechanism is activated.

[0014] Furthermore, the dynamic power mapping strategy is executed based on a pre-established power classification control table.

[0015] Furthermore, it also includes the following steps: Based on the current operating stage and operating condition level, retrieve the corresponding standard speed and torque range from the preset operating stage database; The standard speed and torque are dynamically adjusted based on the power limitation factor to obtain the target output speed and target output torque of the motor. Based on the target output speed and torque of the motor, the drive current of the pump is obtained, and then the control parameters of the pump proportional solenoid valve are set to determine the target absorption torque.

[0016] The technical solution provided by this invention has the following advantages compared with the prior art: This invention applies a hidden Markov model to analyze the main pump pressure signal in real time, automatically and accurately identifying five operation stages: digging, hoisting rotation, unloading, empty bucket return, and digging preparation, providing a precise timing reference for dynamic control. A new three-dimensional collaborative control architecture of "operation stage - real-time load - battery status" is constructed. In the prior art, the motor, hydraulic system, and battery management system are independent, resulting in low collaborative efficiency. This invention deeply integrates the identified operation stage, load condition (light load / normal / heavy load), and real-time state of charge (SOC) of the sodium-ion battery, uniformly acquiring and collaboratively adjusting the motor output and pump absorption torque through a dynamic power mapping strategy, realizing an integrated closed loop of "perception-decision-execution". This not only matches power with load, significantly reducing overflow and idling losses, but also, for the first time, uses the sodium battery SOC as a core constraint, ensuring safety while leveraging its high-rate advantage, achieving global energy efficiency optimization. An adaptive energy management strategy for sodium-ion battery engineering machinery applications is also provided. Existing solutions do not consider the characteristics of sodium batteries. This invention designs a precise SOC estimation and prediction algorithm that incorporates temperature and internal resistance compensation, along with a matching SOC-power grading mapping and dynamic limiting mechanism. This allows for proactive adjustment of output power based on battery status, maximizing battery life and safety while ensuring operational efficiency. Through end-to-end innovation in "intelligent timing identification, multi-source collaborative control, and battery adaptation management," this invention systematically solves the problems of low energy efficiency in traditional excavators and short battery life and slow response in electrified models. Real-world testing shows significant improvements in energy efficiency (15-25%), battery life (over 20%), and unit energy consumption (over 20%), demonstrating outstanding creativity, novelty, and practical value. Attached Figure Description

[0017] Figure 1 This is a flowchart of the HMM model training and verification process provided in an embodiment of the present invention; Figure 2 This is a flowchart of SOC acquisition and power management provided in an embodiment of the present invention; Figure 3 The flowchart for real-time pressure acquisition and feature extraction provided in the embodiments of the present invention is shown. Detailed Implementation

[0018] The following detailed description of a specific embodiment of the present invention is provided in conjunction with the accompanying drawings. However, it should be understood that the scope of protection of the present invention is not limited to the specific embodiment.

[0019] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the technical solution of this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0020] The present invention will be described below through several specific embodiments. To keep the following description of the embodiments clear and concise, detailed descriptions of known functions and components may be omitted. When any component of an embodiment of the present invention appears in more than one drawing, the component may be represented by the same reference numerals in each drawing.

[0021] Figure 1 This is a flowchart of the HMM model training and verification process provided in an embodiment of the present invention; Figure 2 This is a flowchart of SOC acquisition and power management provided in an embodiment of the present invention; Figure 3 The flowchart for real-time pressure acquisition and feature extraction provided in the embodiments of the present invention is shown.

[0022] like Figure 1 and Figure 2As shown, this embodiment of the invention provides a staged power matching control method for a sodium-ion battery excavator, including the following steps: acquiring the pressure signal of the main pump of the hydraulic system in real time; inputting the pressure signal into a hidden Markov model to identify the current operating stage of the excavator in real time, including digging, hoisting and slewing, unloading, empty bucket return, and digging preparation; acquiring the real-time load size under the current operating stage and comparing it with a preset standard load range to determine the current working condition level; acquiring the state of charge (SOC) value of the sodium-ion battery in real time; and obtaining the output power of the sodium-ion battery through a preset dynamic power mapping table based on the current operating stage, working condition level, and SOC value.

[0023] Specifically, a dual-channel pressure sensor is used to acquire the pressure signal of the excavator's main pump. The sensor is installed at the outlet of the excavator's main pump and connected to the data acquisition system via a CAN bus. A dual-channel pressure sensor (range 0-35MPa, accuracy ±0.5%FS) is used, with a sampling frequency of 100Hz to ensure the capture of transient changes. Data processing includes Kalman filtering to eliminate high-frequency noise and acquiring the pressure mean, variance, and rate of change within a 50ms sliding window.

[0024] Furthermore, a Hidden Markov Model (HMM) is a statistical model used to describe a Markov process containing hidden unknown parameters. In excavator operation phase identification, the actual operation phases (such as digging, slewing, unloading, etc.) are hidden, while the dual-pump pressure data are observable. A complete HMM consists of a quintuple λ = (N, M, A, B, π), where N is the number of hidden states; M is the number of observation symbols; A is the state transition matrix; B is the observation probability matrix; and π is the initial state probability vector.

[0025] Furthermore, the pre-trained Hidden Markov Model (HMM) was constructed and trained as follows: historical main pump pressure data of the excavator during multiple complete work cycles were collected, and the data was cleaned and normalized; the processed historical data were manually labeled, and five hidden states were defined to correspond to the five work stages; the pressure data was converted into an observation symbol sequence using an equal-width discretization method; the probability vector π, state transition matrix A, and observation probability matrix B in the initial HMM were iteratively trained using the Baum-Welch algorithm; and the trained model was evaluated using five-fold cross-validation.

[0026] Specifically, the Baum-Welch algorithm is used for unsupervised learning of historical dual-pump pressure sequences. The hidden states include five typical operational stages: digging, hoisting and slewing, unloading, empty bucket return, and digging preparation. The system inputs real-time pressure data into the trained HMM model and outputs the probability distribution of the current operational stage.

[0027] Specifically, the number of hidden states is 5, corresponding to 5 operation stages (S1 excavation, S2 hoisting and rotation, S3 unloading, S4 empty bucket return, S5 excavation preparation); observation symbols: M=100, pressure values ​​are converted into symbol sequences through equal-width discretization; core parameters: initial probability vector π, state transition matrix A, observation probability matrix B.

[0028] Specifically, the Baum-Welch algorithm is an expectation-maximization algorithm that maximizes the likelihood probability of the model with respect to the training data by iteratively updating the initial probability, state transition matrix, and observation probability matrix.

[0029] Furthermore, the pressure sensor sampling frequency is set to 100Hz, continuously collecting more than 100 hours of actual operation data to cover different working conditions and different operators' operating habits.

[0030] The 3σ criterion is used to identify and remove outlier data points. Specifically, the mean μ and standard deviation σ of the pressure data are obtained, and data points exceeding the range [μ-3σ, μ+3σ] are identified as outliers and removed.

[0031] Normalization: The original pressure data is linearly mapped to the [0,1] interval; Work Cycle Annotation: Construct a dataset containing over 10,000 sample segments. Each sample segment represents a complete work cycle, covering five stages: digging, hoisting and slewing, unloading, empty bucket return, and digging preparation. Manually annotate five typical work stages (digging, hoisting and slewing, unloading, empty bucket return, and digging preparation), requiring an annotation accuracy of ≥95%.

[0032] Training and validation process: The Baum-Welch algorithm is used to iteratively optimize parameters, with a convergence condition of likelihood function change rate < 0.001 and a maximum of 1000 iterations. Performance is evaluated using 5-fold cross-validation, requiring a state recognition accuracy ≥ 90%. A regularization term of λ = 0.01 is introduced to prevent overfitting. The dataset is randomly divided into 5 equal parts, and 4 parts are used for training and 1 part for testing, repeated 5 times. The average of the 5 test results is used as the model performance metric to reduce bias caused by random partitioning and improve evaluation reliability.

[0033] Performance enhancement strategies: Parameter optimization: Optimal combination of hidden states and observation symbols in grid search; Feature enhancement: Expanding temporal features such as pressure change rate and pressure variance to improve state discrimination; Evaluation metrics: Comprehensive evaluation using F1-score, recall, and precision (F1-score focuses on balance, recall focuses on false negative rate, and precision focuses on false negative rate). The system achieves accurate identification of operating conditions through a hidden Markov model, providing a decision-making basis for subsequent dynamic energy scheduling.

[0034] Furthermore, it also includes preprocessing the real-time acquired pressure signal. The preprocessing process includes the following steps: performing low-pass digital filtering on the pressure signal; acquiring its time-domain features based on the pressure signal within the sliding window, including the pressure mean, pressure variance, and instantaneous rate of change of pressure; and using the acquired time-domain features as the observation input of the hidden Markov model.

[0035] Furthermore, the digital filtering employs a Butterworth low-pass filter with a cutoff frequency of 20Hz, and the system delay time is controlled within 50ms.

[0036] Specifically, the data acquisition system is configured with a 100Hz sampling frequency (10ms period) and a circular buffer with 1024 sampling points. Real-time data transmission is achieved through a CAN bus to ensure data timeliness.

[0037] Furthermore, the sliding window has a window duration of 50 milliseconds, an overlap rate of 50%, and the temporal features are updated every 50 milliseconds.

[0038] Window configuration: 50ms time window (covering 5 sampling points), 50% overlap, achieving feature updates every 50ms. Temporal feature set:

[0039] Mean pressure: μ = (1 / n)∑Pᵢ; Pressure variance: σ²=(1 / n)∑(Pᵢ-μ)²; Instantaneous rate of change: ΔP / Δt (reflects the dynamic characteristics of pressure); μ(mu): Represents the arithmetic mean of all main pump pressure sampling points within a specific time window. n: Represents the total number of pressure sampling points included in the acquisition window. Pᵢ: Represents the instantaneous pressure signal value acquired at the i-th time point within the window. σ²: Represents the square of the dispersion or fluctuation of the pressure value Pᵢ relative to its average value μ within a time window. (Pᵢ-μ): Represents the deviation of the pressure value at each sampling point from the average pressure within the window.

[0040] Specifically, digital filtering: a second-order Butterworth low-pass filter with a cutoff frequency of 20Hz is used to effectively suppress high-frequency noise, and the system delay is strictly controlled within 50ms; Anomaly detection: Identifies outliers based on the 3σ criterion, repairs outlier data points through linear interpolation, and implements an automatic compensation mechanism for lost data; Integrity verification: Real-time monitoring of data stream integrity and dynamic compensation for missing sampling points.

[0041] Furthermore, the determination of the current operating condition level includes: if the real-time pressure is lower than 80% of the lower limit of the standard load range, it is determined to be a light load condition; if the real-time pressure is between 80% of the lower limit of the standard load range and 120% of the upper limit, it is determined to be a normal operating condition; if the real-time pressure is higher than 120% of the upper limit of the standard load range, it is determined to be a heavy load condition; among these, the current operating condition is finally confirmed only when the operating condition level is consistent for three consecutive sampling periods.

[0042] Furthermore, the process of obtaining the state of charge (SOC) value includes the following steps: establishing a communication connection with the battery management system via the CAN bus; sending an SOC request at a frequency of not less than 10Hz and receiving response data; performing temperature compensation and internal resistance compensation on the obtained SOC value; and using historical SOC values ​​as a substitute if communication fails.

[0043] The remaining battery capacity is obtained based on the battery's state of charge (SOC). The SOC acquisition mechanism uses the CAN 2.0B protocol to communicate with the battery management system (BMS) and transmits data frames (extended frames) at a baud rate of 500kbps. A 300ms timeout retry mechanism is set. The SOC measurement accuracy reaches ±2%, the resolution is 0.1%, and the data update delay is controlled within 100ms.

[0044] Furthermore, it also includes a SOC trend prediction step: using a Kalman filter algorithm to predict the SOC change trend in the next 10 minutes, and if the prediction shows that the SOC decline rate exceeds the threshold, then the power back-off protection mechanism is activated.

[0045] A pre-alarm is triggered when ΔSOC / Δt > 2% / min; the SOC trend for the next 10 minutes is predicted based on Kalman filtering; specifically, parameters such as the "system noise covariance matrix" and "state transition matrix" in Kalman filtering are obtained through experimental calibration. The obtained data is input into the Kalman filtering algorithm model, and Kalman filtering uses a two-step recursive process of prediction and update to optimally estimate SOC and predict its changing trend. Automatic power back-off protection is activated when SOC drops sharply; that is, when the SOC decline rate exceeds a threshold, the motor output power is gradually reduced, limiting the battery discharge current. The degree of back-off can be dynamically adjusted according to the absolute value of SOC and the rate of decline, and is restored after SOC recovers or charging.

[0046] Furthermore, the dynamic power mapping strategy is executed based on a pre-established power classification control table.

[0047] Table 1 Power Level Control Table The state of charge (SOC) output from the BMS is read in real time via the CAN bus to establish an SOC-power mapping strategy. A dynamic compensation mechanism ensures that the target power is automatically reduced by 2% for every 5% decrease in SOC.

[0048] For example, if 75% is within the 70%-90% range, then the State of Charge (SOC) has decreased by 15%. Therefore, the target power reduction is: 0.95 - (0.15 / 0.05) * 0.02 = 0.95 - 0.06 = 0.89. In other words, when the SOC is 75%, the target power is 89%. A SOC of 71% falls within the 70%-90% range, and a SOC of 70% falls within the 50%-70% range. The highest power in this range is 90%, and subsequent reductions are based on this benchmark.

[0049] Furthermore, the process includes the following steps: based on the current operating stage and working condition level, retrieve the corresponding standard speed and torque range from the preset operating stage database; dynamically adjust the standard speed and torque according to the power limitation coefficient to obtain the target output speed and target output torque of the motor; based on the target output speed and torque of the motor, and combined with the hydraulic system efficiency model, obtain the pump drive current, and then set the control parameters of the pump proportional solenoid valve to determine the target absorbed torque.

[0050] The feature extraction process meets real-time requirements, providing temporal dynamic feature support for working condition identification.

[0051] Battery state analysis, temperature compensation: Based on a temperature range of -20°C to 60°C, the SOC value is dynamically corrected through characteristic curves. The characteristic curves were obtained through preliminary experimental calibration in the laboratory.

[0052] Aging compensation: Real-time monitoring of battery internal resistance changes, and correction of estimated values ​​based on the internal resistance-SOC mapping relationship, which is obtained through preliminary experimental calibration in the laboratory.

[0053] Intelligent energy management and adaptive control for excavators: The system adopts a three-layer architecture design: Perception layer: dual-pump pressure sensor, battery management system (BMS); Decision layer: HMM operation phase identification engine, working condition judgment unit, energy optimization control unit; Execution layer: motor controller, hydraulic pump adjustment system. Data support: standard load range database, SOC-power mapping table, historical operation feature library.

[0054] Inference is performed every 200ms. Based on the Hidden Markov Model (HMM), the Viterbi algorithm is used to decode the optimal state sequence and output the most likely work stage P(St|O1:t) at the current moment. A confidence threshold of ≥0.7 is used; states below this threshold are marked as "uncertain states." If a single uncertainty occurs during the process, the current state is inferred from the previous state based on the excavator's work state transitions: "digging, hoisting and slewing, unloading, empty bucket return, digging preparation." If there are multiple consecutive uncertain states, the state is set to "digging."

[0055] Specifically, P(St|O1:t) represents the posterior probability that time t is in state St given all the observation sequences from time 1 to time t; St: the hidden state at time t (i.e., the working stage of the excavator); O1:t: the sequence of all observations from time 1 to time t; P(St|O1:t): the probability of being in a certain state at the current time based on all historical observation data.

[0056] The established "Operation Phase Database" includes parameters such as phase ID, phase name, standard pressure range, standard torque range, standard power range, and average duration. Detailed parameters for each phase can be obtained.

[0057] Operating condition determination: Operating condition is determined based on the comparison between real-time pressure and the standard range during the "excavation" phase: Light load: Real-time pressure < standard lower limit × 0.8, power coefficient 0.75-0.85; Normal: Standard lower limit × 0.8 ≤ real-time pressure ≤ standard upper limit × 1.2, power coefficient 0.95-1.05; Heavy load: Real-time pressure > standard upper limit × 1.2, power coefficient 1.15-1.25 Operating condition stability assessment: Confirmation is only made if the same operating condition is maintained for three consecutive sampling cycles to prevent misjudgment caused by instantaneous pressure fluctuations.

[0058] Motor torque and speed setting: The current motor output speed and torque are obtained based on the current operating conditions and work stage; the final motor output speed and torque are obtained based on the SOC power limit factor; the pump drive current is obtained based on the motor's output speed and torque to set the pump's maximum absorbable torque. Motor control commands are generated and sent to the motor driver via the CAN bus.

[0059] Specifically, through bench tests and field tests, optimal motor speed and torque data are collected under different operating conditions and stages. Based on this, a primary correspondence table of "operating condition - operating stage - theoretical motor speed / torque" is established and stored in a pre-set operating stage database. In actual control, the controller can directly obtain the required theoretical motor output speed and torque by querying this primary correspondence table.

[0060] Through experiments, a second correspondence table was established between the "SOC power limitation factor" and the "motor output speed and torque". Based on the final output speed and torque of the motor, a third correspondence table was established between the "motor output speed and torque" and the "pump target absorbed torque" through pre-calibrated experiments. By consulting these three correspondence tables, the controller can directly set the pump's drive current, thereby achieving precise and rapid control of the pump's maximum absorbed torque.

[0061] The above inventions are merely a few specific embodiments of the present invention. However, the embodiments of the present invention are not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A method for controlling the power matching of a sodium-ion battery excavator in stages, characterized by, The method comprises the following steps: Real-time acquisition of the pressure signal of the main pump of the hydraulic system; Inputting the pressure signal into a hidden Markov model to identify the current working stage of the excavator in real time, wherein the working stage comprises excavation, lifting and rotating, unloading, empty bucket return and excavation preparation; Acquiring the real-time load size under the current working stage and comparing it with the preset standard load range to determine the current working condition level; Real-time acquisition of the state of charge value of the sodium-ion battery; Based on the current working stage, the working condition level and the state of charge value, the output power of the sodium-ion battery is obtained through a preset dynamic power mapping table.

2. The sodium-ion battery excavator phased power matching control method of claim 1, wherein, The pre-trained hidden Markov model is constructed and trained in the following way: Collecting historical main pump pressure data of the excavator in multiple complete working cycles, and performing cleaning and normalization processing on the data; Artificially labeling the processed historical data, and defining five hidden states corresponding to the five working stages respectively; Converting the pressure data into an observation symbol sequence by using the equal-width discretization method; Using the Baum-Welch algorithm to iteratively train the initial probability vector π, the state transition matrix A and the observation probability matrix B in the hidden Markov model; Evaluating the trained model through five-fold cross-validation.

3. The method of claim 1, wherein the method is a method of controlling a sodium-ion battery excavator phased power matching, characterized in that, The method further comprises the following steps of preprocessing the real-time acquired pressure signal: Performing low-pass digital filtering on the pressure signal; Based on the pressure signal in the sliding window, acquiring its time domain features, including pressure mean, pressure variance and pressure instantaneous change rate; Taking the acquired time domain features as the observation input of the hidden Markov model.

4. The method of claim 3, wherein the method further comprises: The digital filtering uses a Butterworth low-pass filter with a cutoff frequency of 20 Hz, and the system delay time is controlled within 50 ms.

5. The method of claim 3, wherein the method further comprises: The window length of the sliding window is 50 ms, and the overlap rate is 50%, and the time domain features are updated every 50 ms.

6. The method of claim 1, wherein the method is a method of controlling a sodium-ion battery excavator phased power matching, characterized in that, The determination of the current working condition level comprises: If the real-time pressure is lower than 80% of the lower limit value of the standard load range, it is determined as light load working condition; If the real-time pressure is between 80% of the lower limit value and 120% of the upper limit value of the standard load range, it is determined as normal working condition; If the real-time pressure is higher than 120% of the upper limit value of the standard load range, it is determined as heavy load working condition; Wherein, only when the working condition level is consistent in three consecutive sampling periods, the current working condition is finally confirmed.

7. The method of claim 1, wherein the method is a method of controlling a sodium-ion battery excavator phased power matching, characterized in that, The process of acquiring the state of charge value comprises the following steps: Establishing a communication connection with the battery management system through the CAN bus; Sending SOC requests and receiving response data at a frequency of not less than 10 Hz; Temperature compensation and internal resistance compensation are performed on the acquired SOC value; If communication fails, the historical SOC value is used as a substitute.

8. The sodium-ion battery excavator phased power matching control method of claim 7, wherein, The method further comprises an SOC trend prediction step: using the Kalman filtering algorithm to predict the SOC change trend in the next 10 minutes, and if the prediction shows that the SOC decline rate exceeds the threshold, a power rollback protection mechanism is started.

9. The method of claim 1, wherein the method is a method of controlling a sodium-ion battery excavator phased power matching, characterized by, The dynamic power mapping strategy is executed based on a pre-established power grading control table.

10. The method of claim 1, wherein the method is a method of controlling a sodium-ion battery excavator phased power matching, characterized in that, The method further comprises the following steps: According to the current working stage and the working condition level, a corresponding standard speed and torque range is called from a preset working stage database; According to the power limitation coefficient, the standard speed and torque are dynamically adjusted to obtain a target output speed and a target output torque of the motor; According to the target output speed and the target output torque of the motor, a driving current of the pump is obtained, and then a control parameter of a pump proportional electromagnetic valve is set to determine a target absorption torque.

Images