Adaptive energy-saving control method and system for bulldozer

Abstract

The application discloses a bulldozer power adaptive energy-saving control method and system, relates to the technical field of engineering machinery power systems, and comprises the following steps: S001, an energy time-frequency baseline and a bidirectional phase mapping model are constructed, a micro-perturbation sequence is injected into an energy interaction chain of an engine and a hydraulic pump, energy coupling response signals are collected, a phase synchronization reference surface is generated, and the dynamic phase difference and energy flow direction of the engine and the hydraulic pump are quantified; S002, a counterfactual playback chain is established based on the phase synchronization reference surface, the energy coupling of a load transition process is time-reversed, a resonance frequency band and a torque valley trigger nucleus are identified, an abnormal energy aggregation path is extracted, and a risk instruction surface is generated. The application constructs an energy phase mapping and self-healing regulation mechanism, realizes quantitative identification and dynamic optimization of energy coupling of the engine and the hydraulic pump, effectively suppresses resonance impact and torque collapse, and improves the stability, energy saving and operation continuity of the bulldozer under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of engineering machinery power system technology, specifically to a bulldozer power adaptive energy-saving control method and system. Background Technology

[0002] Bulldozer power adaptive energy-saving control refers to an energy-saving control strategy that utilizes intelligent sensing and dynamic control technology to automatically adjust engine output power, hydraulic system power supply ratio, and transmission system torque distribution based on real-time operating conditions (such as bulldozing resistance, slope, load changes, engine speed, and hydraulic pressure). Its core idea is to give the bulldozer's power system "adaptive" capabilities, meaning it can automatically match appropriate power output according to actual load demands, rather than continuously operating at high power, thereby reducing ineffective energy consumption and fuel waste. By introducing intelligent algorithms (such as fuzzy control, machine learning, or adaptive PID control), the system can dynamically balance power performance and energy efficiency levels under different terrains, operating conditions, and work intensities, achieving optimal bulldozing efficiency and fuel economy, and significantly extending equipment lifespan and reducing emissions.

[0003] The existing technology has the following shortcomings:

[0004] In existing technologies, bulldozer power systems typically employ an adaptive energy-saving control strategy based on engine speed feedback and hydraulic pump load signals. This strategy reduces fuel consumption by adjusting the engine output power and the hydraulic system's power supply ratio in real time. However, in complex operating scenarios, a phase difference often exists between the dynamic adjustment process of engine speed and the adaptive adjustment process of hydraulic pump load. When the engine's torque output has not yet been adjusted while the hydraulic pump remains under high load, the system's energy transmission chain will experience periodic coupling resonance. This resonance causes a short-term collapse in engine output torque, which the control unit incorrectly identifies as a low-load operating range, triggering a reduction in fuel supply and creating a "false no-load" state. The actual output of the power system deviates significantly from the control expectation. If this mismatch continues to accumulate, it will cause a power collapse in the engine at low speeds, leading to serious consequences such as engine stalling, momentary reverse rotation of the hydraulic pump, or high-pressure oil backflow, resulting in structural damage to the powertrain and operational interruption.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a bulldozer power adaptive energy-saving control method and system to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a bulldozer power adaptive energy-saving control method, comprising the following steps:

[0008] S001, construct an energy time-frequency baseline and a two-way phase mapping model, inject a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collect energy coupling response signals, and generate a phase synchronization reference surface to quantify the dynamic phase difference and energy flow direction between the engine and the hydraulic pump;

[0009] S002, based on the phase synchronization reference surface, establish a counterfactual playback chain, perform time reversal on the energy coupling of the load transition process, identify the resonant frequency band and torque trough trigger core, extract abnormal energy accumulation paths, and generate risk command surface;

[0010] S003, under the risk command surface constraint, train the dynamic forecaster, jointly analyze the injection delay, hydraulic pump swashplate angle response and valve control timing characteristics, generate the prior torque buffer and hydraulic pump suction power boundary, and form a high-dimensional feedforward anchor point.

[0011] S004, perform phase shaping based on high-dimensional feedforward anchor point, inject phase correction sequence and superimpose hysteresis compensation factor, reconstruct engine speed return regulation law, correct energy loop response cycle, cut off resonance positive feedback channel, and freeze pseudo no-load window.

[0012] S005, based on the frozen window, performs energy dual-track diversion control, activates the torque storage chamber and the reverse phase recharge channel, and gates the energy release rhythm according to the residual density parameter to dissipate energy resonance and suppress hydraulic pump reversal, thereby realizing the self-healing closed-loop control of the bulldozer power system.

[0013] Preferably, step S001 includes:

[0014] Speed, torque, injection pulse width, intake pressure, suction flow rate, return pressure, swashplate angle change amplitude, and instantaneous flow pulsation value are collected at the engine output end and hydraulic pump input end, and are sampled synchronously.

[0015] A sequence of small disturbance commands is injected into the engine throttle valve and the hydraulic pump control valve to obtain a response signal;

[0016] The response signal is subjected to bandwidth compression and spectrum expansion to extract disturbance response features and construct a two-way phase mapping relationship;

[0017] The feature nodes in the phase mapping path are fused to establish a phase synchronization reference surface and generate a dynamic energy flow map by combining the energy flow direction calibration signal.

[0018] Preferably, step S002 includes:

[0019] The disturbance triggering region, energy peak region, and stabilization recovery region of the energy response curves of the engine and hydraulic pump are extracted to construct the load transition window and perform time inversion operation.

[0020] The energy response curve after time reversal is mapped to the phase synchronization reference plane, the intersection and separation points are extracted, and the energy resonance seed frequency band is identified.

[0021] Calculate the time lag between the peak output energy and the peak absorption energy, identify the torque trough trigger core, and extract the abnormal energy accumulation path;

[0022] By integrating the energy density, phase offset rate, and hydraulic feedback delay time of path nodes, a risk command surface is generated for control and early warning.

[0023] Preferably, step S003 includes:

[0024] Based on the high-risk response region, the injection delay feature set is extracted to obtain the coupled features of injection start time, combustion delay time and speed response.

[0025] Extract the swashplate angle response characteristics of the hydraulic pump and perform time mapping with the injection delay trigger point to form a correlation description between the swashplate angle suction behavior and fuel input;

[0026] Extract the opening change amplitude and response delay time of the main oil supply control valve and the pressure relief valve to establish a valve control timing feature set;

[0027] By jointly constructing the prior torque buffer and suction power boundary using the injection delay feature set, swashplate angle response feature set, and valve control timing feature set, a high-dimensional feedforward anchor point is generated.

[0028] Preferably, the high-dimensional feedforward anchor point is formed by aligning the injection control time axis, the swashplate angle change path and the valve opening evolution trajectory in time and calibrating the response inflection point, thus creating an energy regulation and control benchmark for adjusting the engine torque release limit and the hydraulic pump suction power boundary in advance.

[0029] Preferably, step S004 includes:

[0030] Extract the injection delay trigger point, the maximum power absorption point of the swashplate angle, and the valve control cutoff point to construct a unified reference surface for dynamic response;

[0031] Based on the benchmark, the actual engine speed response is compared, and a conjugate correction sequence is injected to adjust the speed response trajectory.

[0032] A hysteresis compensation factor is constructed based on the hysteresis region and superimposed on the correction sequence;

[0033] By combining the feedforward anchor point mapping parameters to reconstruct the speed return regulation law, and locking the sham no-load state freeze window after resonance dissipation, the energy loop response structure is stabilized.

[0034] Preferably, the conjugate correction sequence injects adjustment commands at different rhythms during the engine acceleration phase, steady-state maintenance phase, and pullback phase, and smooths the sequence edges through interpolation to achieve continuity and controllability of speed changes.

[0035] Preferably, step S005 includes:

[0036] Based on the frozen window, nonlinear segments in the engine output torque response curve are extracted to construct an energy anomaly accumulation mapping table.

[0037] Establish a fluctuation residual relationship between the peak value of the secondary disturbance that is not leveled by the speed return regulation in the mapping table and the instantaneous load return that cannot be absorbed in the hydraulic pump suction curve;

[0038] A breathing torque energy storage cavity is connected within the torque buffer critical region to achieve on-site energy absorption and buffer resonance disturbances.

[0039] Connect the outlet of the energy storage chamber to the reverse micro-recharge channel and set the energy recirculation trigger condition to complete the controllable recharge process of residual energy.

[0040] Preferably, the inlet of the reverse micro-recharge channel is connected to the outlet of the breathing torque energy storage chamber, and the outlet is connected to the engine intake manifold or the hydraulic pump oil suction buffer chamber. The energy return direction is set by the differential pressure control element, and the channel is dynamically opened according to the synchronicity of the speed fluctuation trend and the phase difference of the hydraulic back pressure curve, so as to realize the delayed recharge and closed-loop reuse of residual energy.

[0041] The bulldozer power adaptive energy-saving control system includes a phase mapping modeling module, an energy playback identification module, a dynamic forecast generation module, a phase shaping correction module, and an energy diversion self-healing module.

[0042] The phase mapping modeling module constructs an energy time-frequency baseline and a two-way phase mapping model, injects a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collects energy coupling response signals, and generates a phase synchronization reference surface to quantify the dynamic phase difference and energy flow direction between the engine and the hydraulic pump.

[0043] The energy replay identification module establishes a counterfactual replay chain based on the phase synchronization reference plane, performs time reversal on the energy coupling of the load transition process, identifies the resonant frequency band and torque trough trigger core, extracts abnormal energy accumulation paths, and generates a risk command plane.

[0044] The dynamic forecast generation module trains the dynamic forecaster under the risk command surface constraint, and jointly analyzes the injection delay, hydraulic pump swashplate angle response and valve control timing characteristics to generate the prior torque buffer and hydraulic pump suction power boundary, forming a high-dimensional feedforward anchor point.

[0045] The phase shaping and correction module performs phase shaping based on the high-dimensional feedforward anchor point, injects the phase correction sequence and superimposes the hysteresis compensation factor, reconstructs the engine speed return regulation law, corrects the energy loop response cycle, cuts off the resonance positive feedback channel, and freezes the pseudo-no-load window.

[0046] The energy diversion self-healing module performs dual-track energy diversion regulation based on the frozen window, activates the torque energy storage chamber and the reverse phase recharge channel, and gates the energy release rhythm according to the residual density parameter to dissipate energy resonance and suppress hydraulic pump reversal, thereby realizing the self-healing closed-loop control of the bulldozer power system.

[0047] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0048] This invention, by constructing a unified energy time-frequency baseline and phase mapping mechanism, achieves for the first time a quantitative expression of the phase difference and energy flow pattern in energy coupling states. Combined with counterfactual playback and high-dimensional anchor point generation mechanisms, it effectively identifies and intervenes in seed frequency bands and torque triggering factors that induce resonance. By introducing conjugate correction and hysteresis compensation mechanisms, it dynamically corrects the energy loop response cycle. Furthermore, by combining a breathing-type torque energy storage structure and energy diversion path, it constructs a self-healing closed-loop control system. This method not only significantly reduces engine speed drop, hydraulic pump reversal, and resonance impact problems caused by energy response mismatch, but also enhances the power system's adaptability to operational disturbances and its energy-saving performance, achieving dual optimization of power performance and energy efficiency, and greatly improving the stability, safety, and operational continuity of bulldozers under dynamic load conditions. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0050] Figure 1 This is a flowchart of the bulldozer power adaptive energy-saving control method of the present invention.

[0051] Figure 2 This is a schematic diagram of the module of the bulldozer power adaptive energy-saving control system of the present invention. Detailed Implementation

[0052] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0053] This invention provides, for example Figure 1 The bulldozer power adaptive energy-saving control method shown includes the following steps:

[0054] S001, Construct a unified energy time-frequency baseline and a two-way phase mapping model, inject a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collect the energy coupling response signal, and generate a phase synchronization reference surface based on the signal to quantify the dynamic phase difference and energy flow pattern between the engine and the hydraulic pump.

[0055] To achieve dynamic identification and phase characteristic characterization of the energy interaction state between the engine and hydraulic pump of a bulldozer under different operating loads, it is first necessary to establish a unified energy time-frequency baseline and construct a two-way phase mapping relationship to form a reference structure that can be used to subsequently determine the energy flow pattern and time-series response state. The specific steps are as follows:

[0056] In the bulldozer power system, the engine output and hydraulic pump input are selected as the two access points of the energy interaction chain. A multi-dimensional real-time data acquisition unit is set up to acquire physical quantities such as engine output speed, torque, injection pulse width, intake pressure, hydraulic pump suction flow rate, return pressure, swashplate angle change amplitude, and instantaneous flow pulsation value at the two access points. Specifically, engine speed and torque are acquired using a high-speed encoder and strain gauge; injection pulse width and intake pressure are acquired via the controller local area bus; hydraulic data are acquired using pressure sensors, flow sensors, and electronically controlled swashplate angle displacement sensors. To avoid data errors, all physical quantities are sampled synchronously with 1ms-level timestamps to ensure strict alignment of cross-domain signals during time-frequency processing. After the above acquisition channels stabilize, a micro-disturbance command sequence with a preset amplitude of 2% to 3% of the rated working load is injected into the engine throttle valve and hydraulic pump control valve, respectively. This disturbance guides the system to generate a weak but observable dynamic response without affecting the stability of the working load, laying the foundation for subsequent analysis.

[0057] Based on the data response results after disturbance injection, the data streams from the engine and hydraulic pump ends are divided into multi-scale time segments. The segmented data streams then undergo bandwidth compression and spectral expansion to extract characteristic frequency bands and their evolution trajectories in the disturbance-triggered response. During this process, a set of associated characteristic curves are constructed by identifying the acceleration response waveform, combustion delay peak, and callback hysteresis trough in the engine response curve, as well as the hydraulic pump suction fluctuation range and back pressure boundary response jitter range. Based on these features, two-directional phase mappings are generated: the first direction maps from engine drive output to hydraulic load changes, and the second direction maps from hydraulic load feedback to engine torque regulation hysteresis. Each direction's mapping starts at the disturbance injection time point, and a multi-order mapping index is established by capturing the energy response inflection points in adjacent intervals after disturbance triggering. This index includes quantitative indicators such as response delay duration, maximum amplitude shift, periodic shift trend, and local steady-state recovery time. These phase mapping relationships are used to establish directional response paths during bidirectional energy coupling.

[0058] After the bidirectional phase mapping relationship is established, the shared feature nodes between the two mapping paths are fused. By matching the time difference, relative amplitude, and phase alignment between the engine output change point and the hydraulic pump load change inflection point, a unified phase synchronization reference surface is established. This reference surface is represented in the form of a two-dimensional time-phase grid and is used to quantify the relative temporal behavior of the engine-side power output and the hydraulic pump-side load response under disturbance triggering. This reference surface covers the complete response cycle before and after the disturbance injection time, and has lateral scalability and longitudinal evolvability, allowing for continuous updates based on new load changes. The generation of the reference surface relies on the common feature index extracted from the previously constructed phase mapping relationship. A continuous and smooth response judgment space is formed by surface fitting through the cumulative lateral phase deviation and the longitudinal response recursive trend. In this space, the phase shift trend of adjacent regions is continuous, and abnormal shift points can form local discontinuous abrupt structures, which can be used to reveal the precursory signs of resonance.

[0059] After establishing a unified phase synchronization reference surface, based on the changes in energy amplitude and phase shift trends in the disturbance response signal, a calibration signal for the energy flow direction between the engine and the hydraulic pump is further superimposed. This, combined with the temporal coupling strength between the engine's fuel injection pulse cycle and the hydraulic pump's suction rhythm, constructs an energy flow direction judgment rule. Specifically, by analyzing the relative position between the rising edge of the fuel injection pulse and the maximum suction point of the hydraulic pump's swashplate angle, the dominant energy direction in the current powertrain is determined. Simultaneously, by combining the changes in swashplate angle rate and the instantaneous torque change of the engine, the active and passive response roles in the energy transfer process are determined. In this way, a dynamic energy flow direction map covering the disturbance cycle is ultimately formed. This map, combined with the phase synchronization reference surface, is used to characterize the dynamic phase difference and energy flow direction relationship during the energy interaction process between the engine and the hydraulic pump, providing an accurate basis for subsequent counterfactual playback and prediction.

[0060] S002, based on the phase synchronization reference surface, establish a counterfactual playback chain, perform time reversal replay of the energy coupling process between the engine and hydraulic pump during load transition, identify the energy resonance seed frequency band and torque trough trigger core, extract the abnormal energy accumulation path, and generate a risk command surface based on the path;

[0061] To accurately analyze the energy coupling mechanism between the engine and hydraulic pump during load transition in the adaptive energy-saving control of bulldozers, it is necessary to perform time reversal and behavior replay of the energy interaction signals based on the previously established phase synchronization reference surface. This is to identify the key frequency bands and triggering factors that lead to energy resonance and torque collapse, and to construct a risk command surface accordingly, enabling early warning and correction of potentially unstable operating conditions. The specific implementation steps are as follows:

[0062] Based on the generated phase synchronization reference surface, the energy response curves of the engine output side and the hydraulic pump load side are segmented and synchronously extracted. A continuous time interval including the disturbance triggering region, the energy peak region, and the stabilization and recovery region is selected to construct a load transition window. This window covers four stages: the rising segment of engine speed fluctuation, the fuel injection delay compensation segment, the hydraulic pump suction acceleration segment, and the return oil pressure adjustment segment. To avoid signal drift and amplitude distortion during time inversion, the original energy signal is normalized in amplitude and calibrated with the phase reference to ensure complete alignment with the phase synchronization reference surface on the time axis. On this basis, the signal flow within the load transition window is arranged in reverse chronological order through a reverse sequence driving method, allowing the energy response paths of the engine and hydraulic pump to be replayed in reverse. The purpose of this time inversion is to recover the dynamic evolution trajectory of the energy chain before resonance occurs, enabling subsequent analysis to identify the antecedent characteristics that lead to system imbalance, rather than relying solely on the final state.

[0063] During signal replay after time inversion, the engine-side energy input curve and the hydraulic pump-side suction response curve are projected onto a unified reference surface based on the phase difference distribution region defined in the phase synchronization reference plane. By tracing the intersection and separation points of the two projected curves along the reverse time direction, the generation area and diffusion path of phase deviation during energy coupling are determined. At each intersection point, the corresponding engine torque change rate, injection pulse width change rate, hydraulic pump flow rate abrupt change amplitude, and swashplate angular acceleration value are extracted, and these parameters are constructed into a time-series feature set to describe the energy interaction intensity at that moment. By continuously tracking the changing trends of these feature sets, a frequency band region where the energy accumulation rate gradually increases and tends to nonlinear diffusion can be identified in the inversion trajectory. This frequency band is the interval where the energy resonance seed frequency band is located. Within this interval, the energy exchange frequencies of the engine and hydraulic pump gradually converge, the coupling phase difference rapidly converges in a short time and then abruptly changes, providing the initial triggering conditions for subsequent torque collapse.

[0064] After identifying the energy resonance seed frequency band, the focus shifted to the energy transfer behavior within this band. By extracting the continuous path of energy flow in the inversion trajectory, the time lag between the peak output energy of the engine and the peak suction power of the hydraulic pump was calculated, and nodes with abrupt changes in lag accompanied by a sharp drop in output torque were identified. Using these nodes as the central region, the combined variation patterns of engine combustion cycle delay time, cylinder pressure rise rate, injection advance angle fine-tuning, and hydraulic pump suction power gradient change rate were analyzed to identify the key trigger point leading to energy imbalance. This trigger point was defined as the torque trough trigger core, which is the source of power collapse and system resonance. To accurately describe the energy evolution characteristics around the trigger core, the abnormal energy accumulation path was extracted by performing symmetrical difference calculations on the energy response curves on both sides of the trigger core. This path reflects the energy channel through which engine power output rapidly migrates to the hydraulic load end on a micro-timescale. Its existence indicates that energy was not effectively absorbed in a short period of time, resulting in a local energy oversaturation zone in the powertrain.

[0065] After identifying the abnormal energy accumulation path, a weighted fusion analysis is performed on the energy accumulation density, phase shift rate, and hydraulic feedback delay time of each node in the path, combined with the energy gradient change trend of each phase region in the phase synchronization reference plane, to form an energy imbalance risk distribution map. Based on this distribution map, a risk command surface is generated in the region where the energy accumulation density exceeds a preset threshold. This command surface uses time and phase as dual coordinate axes to identify the risk intensity range of the engine and hydraulic pump under different phase coupling states. Each risk peak point on the command surface corresponds to a potential resonance triggering stage. When the system operating parameters gradually approach this peak point, the control strategy can perform fuel supply delay adjustment, hydraulic pump suction power limitation, and valve opening slow release operation in advance to avoid energy accumulation forming a resonance link again. Through the above method, the counterfactual playback chain realizes the reconstruction and inversion of the energy coupling history process, enabling the bulldozer power system to perform adaptive prevention and dynamic adjustment based on the historical risk command surface in future operation, thereby providing verifiable energy evolution basis for subsequent dynamic forecasting and phase shaping.

[0066] S003, under the risk command surface constraint, train the cross-domain dynamic forecaster, perform joint analysis on the injection delay characteristics, hydraulic pump swashplate angle response characteristics and valve control timing characteristics, generate the prior torque buffer and hydraulic pump suction power boundary, and construct a high-dimensional feedforward anchor point as the control benchmark for energy regulation.

[0067] To achieve precise energy regulation of the bulldozer's power system under different operating conditions, it is necessary to jointly analyze the dynamic behavior of multiple key control variables based on the high-risk areas marked by the aforementioned risk command surface, thereby constructing a high-dimensional energy regulation benchmark with feedforward regulation capabilities. The specific implementation steps are as follows:

[0068] Based on the established risk command surface, typical torque trough triggering sections and energy anomaly accumulation sections are selected from its high-risk response region. The injection pulse response signal within this region is then meticulously analyzed, and the response time difference between the injection start time, injection duration, combustion delay time, and engine speed jump is extracted as the injection delay feature set. Simultaneously, within the delayed window of the injection pulse change segment, the engine cylinder pressure change rate, intake throttle response curve, and instantaneous torque output at the flywheel edge are monitored to capture the turning point in power response amplitude and combustion stability caused by injection adjustments. During this process, a time series alignment method ensures that the injection signal and torque change behavior are synchronously matched with microsecond-level precision, thereby obtaining a set of feature indices describing the intrinsic coupling relationship between injection action and engine response intensity. These indices not only characterize the precursor effect of the injection strategy on the powertrain but also serve as the input dataset for subsequent dynamic feature association modeling.

[0069] Based on the completed fuel injection delay feature set, the swashplate angle control signal of the hydraulic pump, which completely overlaps with its time period, is selected. The swashplate angle change rate, maximum inclination point time, stable return time, suction range width, and return peak delay time are extracted as swashplate angle response features. Specifically, through the hydraulic pump actuator feedback displacement curve, the swashplate angle displacement waveform is divided into three response intervals: the start-up lifting segment, the steady-state adjustment segment, and the return release segment. The relative time position and response speed of key inflection points in each segment are quantitatively calibrated. Furthermore, by cross-mapping the time span between the swashplate angle peak position and the fuel injection delay trigger point, the active suction tendency and delayed release capability exhibited by the hydraulic pump after fuel injection adjustment can be further identified. This capability directly affects the response efficiency of the hydraulic circuit to power changes. Based on this mapping result, a dependent description of swashplate angle behavior on the fuel input strategy is formed, thereby revealing the time-dependent characteristics and inertial compensation capability of the hydraulic unit in energy fluctuation control.

[0070] After completing the injection delay feature set and swashplate angle response feature set, the corresponding valve control signal behavior features are further extracted. This feature set consists of four dimensions: the opening change amplitude, response delay time, cut-off holding time, and restart inertia time of the main fuel supply control valve, pressure relief valve, return oil slow-release valve, and directional control valve. By bidirectionally mapping the signal change interval of each control valve in the high-risk window of the risk command plane to the injection-swashplate angle combined feature area, the opening and closing rhythm of each valve and the regulation behavior of the main power chain are analyzed synchronously, thereby obtaining a set of time-series control features describing the response hysteresis and energy on / off logic of the control valves. This feature set reflects the key role of the coordinated action of each control valve in avoiding energy concentration and reverse impact during complex load transitions, and reveals the harmonizing performance of the hydraulic control link on the power source output rhythm.

[0071] Based on the three feature sets mentioned above, a cross-dimensional joint behavioral feature space is constructed. Through methods such as time node alignment, trend synchronization analysis, and hysteresis response weight evaluation, the coupling strength between injection delay, swashplate angle response, and valve control timing is determined. Within this joint feature space, a priori torque buffer is constructed by identifying the co-occurring timing inflection points of the three feature sets. This allows for the early determination of the engine's acceptable torque release limit before an impending power collapse or high-load impact is predicted. Simultaneously, based on the upper limit of the swashplate angle increase rate and the changing trend of the hydraulic suction zone length, the hydraulic pump's suction boundary range is determined as a hard constraint condition for hydraulic unit load adjustment. The priori torque buffer and suction boundary parameters are combined and mapped onto the injection control time axis, swashplate angle change path, and valve control opening evolution trajectory to jointly construct a high-dimensional feedforward anchor point. This anchor point not only characterizes the reference response coordinates during energy regulation but also serves as a control benchmark for subsequent phase shaping control and energy release rhythm optimization, enabling proactive energy scheduling and regulation of the bulldozer power system under complex operating conditions.

[0072] S004, based on the high-dimensional feedforward anchor point, performs phase shaping operation, injects phase conjugate correction sequence and superimposes hysteresis compensation factor, reconstructs engine speed return regulation law, corrects the response cycle of energy loop, cuts off the positive feedback channel in energy resonance chain, and realizes the freezing of pseudo no-load state window;

[0073] To achieve dynamic optimization control of the energy loop response rhythm of the bulldozer power system under complex load conditions, it is necessary to perform cross-domain phase shaping operations based on the constructed high-dimensional feedforward anchor point. By introducing a phase correction mechanism and dynamic compensation strategy, the engine speed response behavior is adjusted, eliminating the self-excited channel in the energy resonance chain, thereby achieving response window freezing under sham no-load conditions and improving system stability. The specific implementation steps are as follows:

[0074] Based on the key control parameters defined in the high-dimensional feedforward anchor point, the relative positions of the corresponding injection delay trigger point, the maximum power absorption point of the hydraulic pump swashplate angle, and the valve control cutoff point on the time axis are extracted to construct a unified reference surface for dynamic response. This reference surface serves as the initial reference structure for phase shaping, constrained by the energy response lag trend during high-risk periods, and selects the engine speed change rate, torque release slope, and speed recovery delay time as the main control indicators. Based on this, the envelope of the original speed response curve is constructed. By comparing the actual response curve under the current load condition with the theoretical optimal curve in the feedforward anchor point, the phase shift amplitude and period delay degree are determined. To achieve the phase correction target, a set of conjugate correction sequences based on a periodic reversal configuration is injected into the original speed command. This sequence consists of multiple time segments, acting on the engine acceleration phase, steady-state maintenance phase, and recovery phase, respectively, aiming to pull the engine speed response rhythm back to the desired trajectory corresponding to the anchor point. During the injection of the correction sequence, to avoid mechanical shock caused by command step jumps, interpolation transitions are used to smooth the initial edge, ensuring continuous and controllable speed changes.

[0075] After establishing the conjugate correction sequence, the response lag region in the high-dimensional feedforward anchor point is further identified. The maximum lag time of the injection response, the advance time of the hydraulic pump suction front, and the lag angle displacement of the valve control response are extracted as hysteresis features, and a hysteresis compensation factor sequence is constructed based on these features. This factor sequence is coupled and superimposed with the conjugate correction sequence on the time axis, and applies different weights to the speed command for different control stages, forming a set of dynamic hysteresis correction control paths. In the early stage of engine torque release, the hysteresis compensation factor acts in a peak manner to counteract the sudden drop in speed caused by the hydraulic load's advance suction. In the steady-state maintenance stage, the compensation factor is continuously applied in an average manner to stabilize the engine output frequency. In the pullback stage, the compensation factor is released in a decreasing trend, allowing the speed to gradually transition to the preset minimum control threshold. In this way, the hysteresis compensation factor and the conjugate correction sequence are integrated to achieve dynamic adjustment of the engine output path throughout the entire process, ensuring that the power response is not affected by energy chain lag factors, thereby improving the accuracy and stability of control.

[0076] After the joint control structure is established, the engine speed recovery regulation is reconstructed based on its output. This regulation uses a high-dimensional feedforward anchor point as the core control basis, mapping the torque buffer, hydraulic pump suction boundary, and control valve response time window defined in the feedforward anchor point onto the engine speed control trajectory. In specific implementation, firstly, each parameter is projected onto a unified time axis, and response time sequence segments are constructed. Secondly, a piecewise response model is constructed based on the load characteristics of each segment. For example, in the maximum suction stage, the speed regulation is designed as a gradual increase-maintenance-gradient mode to provide maximum stable output; while in the energy recovery stage, it is set as a steep decrease-suppression-gradient recovery mode to quickly cut off the unstable energy loop. Based on the above, the conjugate correction sequence and hysteresis compensation factor are implanted into each stage to achieve dynamic adjustment of the speed curve. In the entire regulation construction process, special attention is paid to the time coordination relationship between engine speed and hydraulic pump swashplate angle. By introducing a fine-tuning window at the phase synchronization point, it is ensured that the two achieve synchronous resonance avoidance in the critical response stage, thereby destroying the original positive feedback coupling structure.

[0077] After the speed return regulation is executed, the resonance chain in the energy loop is verified. By analyzing the synchronous changes of the torque output curve and the hydraulic pump suction response curve after speed regulation, it is determined whether a periodic resonance trend still exists. If the resonance behavior is confirmed to have dissipated, the corresponding time period is designated as a "false no-load" state freeze window. In this window, the system maintains the current speed output structure unchanged, while rigidly limiting the pressure relief rate and swashplate return speed in the hydraulic loop to prevent the energy return path from reopening. Through the freeze window mechanism, the power control unit can be effectively prevented from responding to short-term low-load signals with fuel reduction, thereby avoiding torque misjudgment and power collapse. Thus, by using a phase shaping control strategy based on a high-dimensional feedforward anchor point, not only is the engine speed response rhythm reconstructed, but the chain-breaking control of the false no-load induced mechanism is also achieved, providing a forward-looking guarantee for the stable operation of the entire power system.

[0078] S005, based on the frozen window, performs singular value-driven dual-track energy diversion regulation, activates the breathing torque energy storage chamber and the reverse micro-recharge channel, dynamically gates the energy release rhythm according to the residual density parameter, completes energy resonance dissipation and hydraulic pump reverse reversal suppression, and forms the self-healing control of the bulldozer power system.

[0079] To further extend the phase-shaping control effect within the frozen window and to actively dissipate the residual energy resonance effect within the power system and prevent and control hydraulic reverse impact, an energy diversion mechanism needs to be introduced based on the frozen window. This mechanism, combined with a specific torque absorption structure and a micro-energy feedback channel, dynamically constructs a closed-loop self-healing energy control process. The specific steps are as follows:

[0080] Based on the high-risk phase coupling segments identified within the frozen window and their corresponding speed return regulation, the nonlinear segments of the engine output torque response curve are extracted, and an energy anomaly accumulation mapping table is constructed. This mapping table uses the secondary disturbance peaks in the engine output waveform that are not successfully leveled by the speed regulation as input, and the instantaneous load return amount that cannot be absorbed by the hydraulic pump suction curve as output, establishing a fluctuation residual relationship between the two. Subsequently, a dynamic response anomaly monitoring device is used to perform high-frequency detection of the rate of change of acceleration of the engine flywheel outer ring to obtain the time window in which the disturbance residuals are concentrated. Within this time window, the high-fluctuation interval representing the secondary resonance source is marked as the torque buffer critical region. To achieve on-site energy absorption and instantaneous dissipation, a breathing torque energy storage cavity structure is injected into this buffer critical region. This energy storage cavity, through a combination design of a variable-capacity cavity and a high-pressure airbag, possesses the capabilities of torque peak absorption, energy hysteresis release, and back pressure buffering. Its inlet and outlet are connected to the flywheel end output shaft and the hydraulic main pipeline, respectively, and are equipped with a mechanically triggered one-way linkage valve to ensure that when the engine torque pulse amplitude exceeds the upper limit of the anchor point stability value, the energy is automatically diverted to the energy storage chamber, avoiding all of it being transferred to the hydraulic circuit and causing backflow.

[0081] While the energy storage chamber completes primary absorption, a reverse-phase micro-recharge channel is designed to further recover and reuse residual energy. This channel uses a bidirectional damping valve to control the return flow of small-scale energy to the main power path and introduces an adjustable delay structure to allow for manually set recharge triggering timing. The channel inlet is connected to the energy storage chamber outlet or the high-pressure hydraulic branch, and the outlet is connected to the engine intake manifold or the hydraulic pump suction buffer chamber. Unidirectional energy introduction is achieved by setting a differential pressure control element. In specific implementation, a set of synchronous opening conditions based on the phase difference between the speed fluctuation trend and the hydraulic back pressure curve is set for this channel. When the above two indicators synchronously approach zero slope and remain within the stable range of the dynamic anchor point, the channel will automatically open, allowing low-density residual energy to recharge the main path. In this way, not only is the bidirectional diversion of the energy path completed, but the remaining resonance factors in the main energy chain are also gradually guided out of the feedback closed loop, thereby physically disconnecting the resonance closed-loop coupling path.

[0082] After achieving physical energy diversion and path transfer, a dynamic gating mechanism is needed to enhance the adaptability and adjustability of the energy release process. This mechanism expands upon the residual density parameter set constructed in the previous stage. Before the freeze window ends, it calculates the energy release offset value per unit time by integrating the difference between the engine speed change curve and the hydraulic pump flow curve. This offset value represents the degree of mismatch in the current energy regulation effect. Based on this value, the current energy state is divided into three response states: high residual release zone, medium residual stable zone, and low residual recharge zone. Energy release thresholds are set for different states, and a step-by-step response strategy is implemented through proportional pressure valve groups. For example, in the high residual release zone, the proportional valve is fully open to accelerate the energy outward conduction speed; in the medium residual stable zone, the proportional valve remains partially open for continuous and stable regulation; and in the low residual recharge zone, the proportional valve is partially closed and the recharge channel is opened, slowly introducing the remaining energy into the main path. This gating method has the ability to adjust the energy release path and release duration in real time, ensuring that the power system can automatically identify the optimal energy distribution path and achieve precise control even under different operating intensities and terrain changes.

[0083] Under the synergistic effect of various energy regulation structures, a set of indicators for identifying the resonance characteristics of the power system is established by analyzing the synchronicity among engine speed change trends, hydraulic pump power absorption trends, and energy residual release frequency. When this set of indicators remains stable over a certain period of time, it can be determined that the energy resonance has been effectively dissipated. At this point, by locking this data segment as a marker of structural self-healing completion, this marker is used as a reference before the next load transition prediction to guide the selection of initial conditions for subsequent fuel injection rhythm, swashplate angle adjustment rate, and speed rhythm planning. In addition, the final energy release residual regression value is compared with the initial residual value of the freeze window. When the difference between the two is less than a preset threshold, the freeze window closure command is triggered, officially ending the energy regulation process for this round. In this way, the bulldozer power system completes the dynamic evolution control of the entire process from energy resonance formation, identification, response, diversion, dissipation to structural closed-loop self-healing, significantly improving energy efficiency and disturbance rejection capability while ensuring system stability.

[0084] This invention, by constructing a unified energy time-frequency baseline and phase mapping mechanism, achieves for the first time a quantitative expression of the phase difference and energy flow pattern in energy coupling states. Combined with counterfactual playback and high-dimensional anchor point generation mechanisms, it effectively identifies and intervenes in seed frequency bands and torque triggering factors that induce resonance. By introducing conjugate correction and hysteresis compensation mechanisms, it dynamically corrects the energy loop response cycle. Furthermore, by combining a breathing-type torque energy storage structure and energy diversion path, it constructs a self-healing closed-loop control system. This method not only significantly reduces engine speed drop, hydraulic pump reversal, and resonance impact problems caused by energy response mismatch, but also enhances the power system's adaptability to operational disturbances and its energy-saving performance, achieving dual optimization of power performance and energy efficiency, and greatly improving the stability, safety, and operational continuity of bulldozers under dynamic load conditions.

[0085] This invention provides, for example Figure 2 The bulldozer power adaptive energy-saving control system shown includes a phase mapping modeling module, an energy playback identification module, a dynamic forecast generation module, a phase shaping correction module, and an energy diversion self-healing module.

[0086] The phase mapping modeling module constructs an energy time-frequency baseline and a two-way phase mapping model, injects a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collects energy coupling response signals, and generates a phase synchronization reference surface to quantify the dynamic phase difference and energy flow direction between the engine and the hydraulic pump.

[0087] The energy replay identification module establishes a counterfactual replay chain based on the phase synchronization reference plane, performs time reversal on the energy coupling of the load transition process, identifies the resonant frequency band and torque trough trigger core, extracts abnormal energy accumulation paths, and generates a risk command plane.

[0088] The dynamic forecast generation module trains the dynamic forecaster under the risk command surface constraint, and jointly analyzes the injection delay, hydraulic pump swashplate angle response and valve control timing characteristics to generate the prior torque buffer and hydraulic pump suction power boundary, forming a high-dimensional feedforward anchor point.

[0089] The phase shaping and correction module performs phase shaping based on the high-dimensional feedforward anchor point, injects the phase correction sequence and superimposes the hysteresis compensation factor, reconstructs the engine speed return regulation law, corrects the energy loop response cycle, cuts off the resonance positive feedback channel, and freezes the pseudo-no-load window.

[0090] The energy diversion self-healing module performs dual-track energy diversion regulation based on the frozen window, activates the torque energy storage chamber and the reverse phase recharge channel, and gates the energy release rhythm according to the residual density parameter to dissipate energy resonance and suppress hydraulic pump reversal, thereby realizing the self-healing closed-loop control of the bulldozer power system.

[0091] The bulldozer power adaptive energy-saving control method provided in this embodiment of the invention is implemented through the above-mentioned bulldozer power adaptive energy-saving control system. For details of the specific methods and processes of the bulldozer power adaptive energy-saving control system, please refer to the above-mentioned embodiment of the bulldozer power adaptive energy-saving control method, which will not be repeated here.

[0092] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A bulldozer power adaptive energy-saving control method, characterized in that, Includes the following steps: S001, construct an energy time-frequency baseline and a two-way phase mapping model, inject a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collect energy coupling response signals, and generate a phase synchronization reference surface to quantify the dynamic phase difference and energy flow direction between the engine and the hydraulic pump; S002, based on the phase synchronization reference surface, establish a counterfactual playback chain, perform time reversal on the energy coupling of the load transition process, identify the resonant frequency band and torque trough trigger core, extract abnormal energy accumulation paths, and generate risk command surface; S003, under the risk command surface constraint, train the dynamic forecaster, jointly analyze the injection delay, hydraulic pump swashplate angle response and valve control timing characteristics, generate the prior torque buffer and hydraulic pump suction power boundary, and form a high-dimensional feedforward anchor point. S004, perform phase shaping based on high-dimensional feedforward anchor point, inject phase correction sequence and superimpose hysteresis compensation factor, reconstruct engine speed return regulation law, correct energy loop response cycle, cut off resonance positive feedback channel, and freeze pseudo no-load window. S005, based on the frozen window, performs dual-track energy diversion control, activates the torque storage chamber and the reverse-phase recharge channel, and gates the energy release rhythm according to the residual density parameter to dissipate energy resonance and suppress hydraulic pump reversal.

2. The bulldozer power adaptive energy-saving control method according to claim 1, characterized in that, Step S001 includes: Speed, torque, injection pulse width, intake pressure, suction flow rate, return pressure, swashplate angle change amplitude, and instantaneous flow pulsation value are collected at the engine output end and hydraulic pump input end, and are sampled synchronously. A sequence of small disturbance commands is injected into the engine throttle valve and the hydraulic pump control valve to obtain a response signal; The response signal is subjected to bandwidth compression and spectrum expansion processing to extract disturbance response features and construct a two-way phase mapping relationship; The feature nodes in the phase mapping path are fused to establish a phase synchronization reference surface and generate a dynamic energy flow map by combining the energy flow direction calibration signal.

3. The bulldozer power adaptive energy-saving control method according to claim 1, characterized in that, Step S002 includes: The disturbance triggering region, energy peak region, and stabilization recovery region of the energy response curves of the engine and hydraulic pump are extracted to construct the load transition window and perform time inversion operation. The energy response curve after time reversal is mapped to the phase synchronization reference plane, the intersection and separation points are extracted, and the energy resonance seed frequency band is identified. Calculate the time lag between the peak output energy and the peak absorption energy, identify the torque trough trigger core, and extract the abnormal energy accumulation path; By integrating the energy density, phase offset rate, and hydraulic feedback delay time of path nodes, a risk command surface is generated for control and early warning.

4. The bulldozer power adaptive energy-saving control method according to claim 1, characterized in that, Step S003 includes: Based on the high-risk response region, the injection delay feature set is extracted to obtain the coupled features of injection start time, combustion delay time and speed response. Extract the swashplate angle response characteristics of the hydraulic pump and perform time mapping with the injection delay trigger point to form a correlation description between the swashplate angle suction behavior and fuel input; Extract the opening change amplitude and response delay time of the main oil supply control valve and the pressure relief valve to establish a valve control timing feature set; By jointly constructing the prior torque buffer and suction power boundary using the injection delay feature set, swashplate angle response feature set, and valve control timing feature set, a high-dimensional feedforward anchor point is generated.

5. The bulldozer power adaptive energy-saving control method according to claim 4, characterized in that, The high-dimensional feedforward anchor point aligns the injection control time axis, swashplate angle change path, and valve opening evolution trajectory in time and calibrates the response inflection point to form an energy regulation control benchmark for adjusting the engine torque release limit and hydraulic pump suction power boundary in advance.

6. The bulldozer power adaptive energy-saving control method according to claim 1, characterized in that, Step S004 includes: Extract the injection delay trigger point, the maximum power absorption point of the swashplate angle, and the valve control cutoff point to construct a unified reference surface for dynamic response; Based on the benchmark, the actual engine speed response is compared, and a conjugate correction sequence is injected to adjust the speed response trajectory. A hysteresis compensation factor is constructed based on the hysteresis region and superimposed on the correction sequence; By combining the feedforward anchor point mapping parameters to reconstruct the speed return regulation law, and locking the sham no-load state freeze window after resonance dissipation, the energy loop response structure is stabilized.

7. The bulldozer power adaptive energy-saving control method according to claim 6, characterized in that, The conjugate correction sequence injects adjustment commands at different rhythms during the engine acceleration phase, steady-state maintenance phase, and pullback phase, and smooths the sequence edges through interpolation to achieve continuity and controllability of speed changes.

8. The bulldozer power adaptive energy-saving control method according to claim 1, characterized in that, Step S005 includes: Based on the frozen window, nonlinear segments in the engine output torque response curve are extracted to construct an energy anomaly accumulation mapping table. Establish a fluctuation residual relationship between the peak value of the secondary disturbance that is not leveled by the speed return regulation in the mapping table and the instantaneous load return that cannot be absorbed in the hydraulic pump suction curve; A breathing torque energy storage cavity is connected within the torque buffer critical region to achieve on-site energy absorption and buffer resonance disturbances. Connect the outlet of the energy storage chamber to the reverse micro-recharge channel and set the energy recirculation trigger condition to complete the controllable recharge process of residual energy.

9. The bulldozer power adaptive energy-saving control method according to claim 8, characterized in that, The inlet of the reverse micro-recharge channel is connected to the outlet of the breathing torque energy storage chamber, and the outlet is connected to the engine intake manifold or the hydraulic pump oil suction buffer chamber. The energy return direction is set by the differential pressure control element, and the channel is dynamically opened according to the synchronicity of the speed fluctuation trend and the phase difference of the hydraulic back pressure curve.

10. A bulldozer power adaptive energy-saving control system, used to implement the bulldozer power adaptive energy-saving control method according to any one of claims 1-9, characterized in that, It includes a phase mapping modeling module, an energy playback identification module, a dynamic forecast generation module, a phase shaping and correction module, and an energy shunting self-healing module: The phase mapping modeling module constructs an energy time-frequency baseline and a two-way phase mapping model, injects a micro-amplitude perturbation sequence into the energy interaction chain between the engine and the hydraulic pump, collects energy coupling response signals, and generates a phase synchronization reference surface to quantify the dynamic phase difference and energy flow direction between the engine and the hydraulic pump. The energy replay identification module establishes a counterfactual replay chain based on the phase synchronization reference plane, performs time reversal on the energy coupling of the load transition process, identifies the resonant frequency band and torque trough trigger core, extracts abnormal energy accumulation paths, and generates a risk command plane. The dynamic forecast generation module trains the dynamic forecaster under the risk command surface constraint, and jointly analyzes the injection delay, hydraulic pump swashplate angle response and valve control timing characteristics to generate the prior torque buffer and hydraulic pump suction power boundary, forming a high-dimensional feedforward anchor point. The phase shaping and correction module performs phase shaping based on the high-dimensional feedforward anchor point, injects the phase correction sequence and superimposes the hysteresis compensation factor, reconstructs the engine speed return regulation law, corrects the energy loop response cycle, cuts off the resonance positive feedback channel, and freezes the pseudo-no-load window. The energy diversion self-healing module performs dual-track energy diversion regulation based on the frozen window, activates the torque energy storage chamber and the reverse phase recharge channel, and gates the energy release rhythm according to the residual density parameter to dissipate energy resonance and suppress hydraulic pump reversal.

Images