Hybrid operation platform power distribution method suitable for narrow lane operation

Abstract

The invention relates to the technical field of hybrid power system power distribution, and provides a hybrid power operation platform power distribution method suitable for narrow lane operation. A state acquisition module, a fuzzy preprocessing unit, a particle filter estimation unit, a model prediction control unit and a self-adaptive control unit are designed, vehicle inclination angle, load and wheel speed data are acquired in real time, and terrain uncertainty factors are output through fuzzy logic processing; estimating hidden state variables such as the wheel slippage rate and the energy margin by adopting a particle filtering algorithm; performing multi-stage power distribution optimization by applying a model predictive control technology to realize coordinated control of the diesel engine and the electric pump; and finally, self-adaptive feedback and switching are executed, wherein the power modes are switched according to the uncertainty factors, inclination angle and load safety protection is implemented, and braking energy is recycled. The terrain adaptability, the power distribution precision and the energy utilization efficiency of the working platform in the narrow lane complex environment are effectively improved, and meanwhile the operation stability and the environment friendliness are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hybrid power system power distribution, in particular to a hybrid working platform power distribution method suitable for narrow lane operation. BACKGROUND

[0002] With the continuous advancement of urbanization and the increasing demand for infrastructure maintenance, the safe operation and reliable power supply of the power system have become a key factor to ensure the normal operation of the city. In the daily maintenance and repair work of urban power distribution network, the aerial work platform as an important special equipment undertakes key tasks such as overhead line maintenance, equipment replacement, fault diagnosis, etc. Especially in the narrow streets and lanes of old urban areas, due to the special topographic conditions such as dense historical buildings, narrow roads and small turning radius, higher requirements are put forward for the maneuverability, flexibility and environmental adaptability of the working platform. Traditional large working vehicles are difficult to enter these areas, while small working platforms face the problems of insufficient power and low working efficiency, etc., so it is urgent to develop new power systems and control methods suitable for narrow lane operation. Hybrid power technology, as an advanced power system combining the advantages of internal combustion engine and electric motor, has been widely used in the fields of automobiles, engineering machinery, etc. in recent years. Hybrid power system can flexibly allocate power output through the cooperative work of diesel engine and electric motor in different working conditions, which can not only take advantage of the large power and strong endurance of diesel engine, but also utilize the fast response, high control precision and zero emission characteristics of electric motor, so as to realize the efficient utilization of energy and environmentally friendly operation. However, the existing hybrid power system is mostly applied to highway vehicles or general engineering machinery, and its power distribution strategy is mainly designed for flat road or regular working conditions, which is not suitable for the special application scene of narrow lane operation.

[0003] In the narrow lane working environment, the working platform faces special challenges such as complex terrain, uncertain road conditions, and limited space. The ground may have slope changes, uneven road surfaces, and slippery conditions, and the vehicle inclination and tire slip state may change at any time, which puts higher requirements on the power distribution strategy. At the same time, the narrow lane environment is usually located in residential or commercial areas, with strict restrictions on noise and emissions, requiring the working platform to minimize the impact on the surrounding environment while ensuring operational capability. In addition, when the working platform is performing high-altitude operations, the extension of the boom and the change in load will cause the center of gravity of the vehicle to shift, threatening the stability of the vehicle, and the power system needs to be able to respond in real time and take appropriate protective measures. The existing hybrid working platform power distribution methods mainly have the following technical problems. First, the traditional rule-based power distribution strategy uses simple threshold judgment and fixed proportional allocation, which cannot effectively handle the terrain uncertainty in the narrow lane environment. Traditional methods usually control in segments according to the size of power demand, using fixed diesel power and electric power proportion in different power intervals, lacking the ability to quantitatively evaluate and adaptively adjust the complexity of the terrain. When the working platform encounters complex conditions such as slopes, bumpy roads, or load mutations in the narrow lane, the fixed allocation ratio cannot be adjusted flexibly according to the actual situation, resulting in mismatched power output, which may either result in insufficient power affecting operational efficiency or excessive power causing energy waste. Second, existing methods have insufficient accuracy in hidden state estimation. Key state variables such as tire slip rate and energy surplus are difficult to measure directly and need to be estimated through filtering algorithms. The traditional extended Kalman filter method assumes that the system model is linear and the noise follows a Gaussian distribution, which is often not true in the narrow lane complex terrain. When the vehicle is driving on uneven roads, the interaction between the tire and the ground exhibits strong nonlinear characteristics, and using linearization methods for state estimation will produce large errors, resulting in inaccurate judgments of the actual running state of the vehicle. Power distribution decisions based on inaccurate state information will inevitably deviate from the optimal solution, affecting the overall performance of the system. Simple moving average filtering has a small computational load, but its filtering effect is limited when facing sensor noise and model uncertainty, and it cannot meet the demand for accurate control. Third, existing energy management strategies lack foresight and global optimization capabilities. Rule-based methods are essentially an immediate response strategy that determines power distribution based on the current state and demand, and cannot predict future conditions. When the working platform encounters consecutive complex conditions such as uphill followed by boom extension operations, control systems that lack predictive capabilities cannot prepare in advance, resulting in delayed power response. In addition, these methods usually only consider meeting the current power demand, without further optimizing energy consumption, stability, and other multiple targets under the premise of meeting demand, resulting in low energy utilization efficiency. Frequent switching of power distribution also causes dramatic fluctuations in power output, increasing wear and tear on mechanical components and reducing ride comfort.Fourth, the existing system generally lacks an effective energy recovery mechanism. During the operation process, a large amount of potential energy and kinetic energy will be generated when the boom is retracted and the vehicle is braked, which will be dissipated in the form of heat if not recovered, resulting in energy waste. Although some hybrid systems have regenerative braking function, its application on special operation platform is not mature enough, especially the energy recovery and utilization technology of hydraulic system is still in the exploratory stage. Low energy recovery efficiency directly leads to limited system endurance, increasing the frequency of external charging and dependence on energy supply.

[0004] Based on the problems existing in the prior art, it is urgent to develop an intelligent hybrid power distribution method suitable for narrow lane operation, and to comprehensively improve the performance and environmental adaptability of the narrow lane operation platform. SUMMARY

[0005] To solve the problems in the background art, the present application provides a hybrid operation platform power distribution method for narrow lane operation, comprising the following steps: S1: real-time state acquisition and fuzzy preprocessing: acquiring vehicle inclination, acceleration, working bucket load and four wheel speed data; dividing the inclination data into three fuzzy sets of low, medium and high, dividing the load data into three fuzzy sets of light, medium and heavy, and dividing the speed data into three fuzzy sets of slow, medium and fast; processing three input fuzzy sets through a fuzzy logic controller to output an uncertainty factor ; S2: particle filter hidden state estimation: initializing a particle swarm; based on the collected sensor data, estimating the wheel slip rate and energy surplus through particle prediction, weight update and resampling process; S3: MPC multi-level prediction distribution: setting the prediction time domain and sampling time; constructing a target function including energy consumption term, inclination term and slip rate term; setting power balance constraint and battery power constraint; solving the quadratic programming optimization problem to output primary diesel power distribution, secondary electric power distribution and tertiary hydraulic peak buffer distribution; S4: adaptive feedback and switching: executing feedback control; judging whether the uncertainty factor is greater than the threshold value, if greater than the threshold value, switching to the conservative mode to increase the proportion of electric power; judging whether the inclination and load are out of limit, if out of limit, triggering the protection limit boom action; starting energy recovery to charge the battery when braking.

[0006] Further, the S1 specifically comprises: S11: acquiring the front and rear inclination and left and right inclination of the vehicle through the IMU sensor at a frequency of 100Hz, in degrees; when or , setting the inclination alarm flag to 1; S12: acquiring the current load weight of the working bucket through the load sensor at a frequency of 50Hz. , unit: kg; the front left wheel speed , the front right wheel speed , the rear left wheel speed , the rear right wheel speed , unit: rpm; S13: according to the collected inclination , , define the membership function: when , belong to the "low" fuzzy set, when , belong to the "medium" fuzzy set, when , belong to the "high" fuzzy set; according to the load , define the membership function: when , belong to the "light" fuzzy set, when , belong to the "medium" fuzzy set, when , belong to the "heavy" fuzzy set; according to the average speed , define the membership function: when , belong to the "slow" fuzzy set, when , belong to the "medium" fuzzy set, when , belong to the "fast" fuzzy set; wherein is the inclination value; is the load value; is the average speed; , , , , respectively, the four wheel speeds; S14: establish a fuzzy rule base, containing 27 rules; input the membership values of the three fuzzy sets, solve the fuzzy by the barycentric method, calculate the output uncertainty factor ; wherein is the activation degree of the rule; is the output value corresponding to the rule; represents the summation operator; convert the fuzzy terrain information into a quantitative uncertainty factor.

[0007] Further, the S2 specifically comprises: S21: initialize particles, the initial state vector of each particle , wherein is the initial wheel slip rate, is the initial diesel engine speed, is the initial electric pump power, is the initial battery power; the initial weight of each particle is set to ; wherein is the total number of particles; is the particle index; superscript denotes transpose; S22: performing state prediction for each particle, the prediction formula is ; wherein is the predicted state of the th particle at the th time; is a nonlinear state transition function; is the control input vector at the th time, including the diesel engine throttle opening and the electric pump start-stop signal; is the process noise, which is subject to Gaussian distribution with mean 0 and variance ; S23: obtaining sensor observation ; calculating the likelihood probability of each particle ; updating the weight of each particle ; normalizing the weight ; wherein is the observation vector at the th time; is the observed average wheel speed; is the observed battery current; is the observation function; is the observation noise covariance; denotes the Euclidean norm of a vector; is the natural exponential function; is the summation index; S24: calculating the number of effective particles ; when , resampling is performed: resample the particles according to the size of the weight , retain high-weight particles and eliminate low-weight particles; after resampling, reset all weights to ; calculating the wheel slip rate estimate and the energy surplus estimate ; wherein is the number of effective particles; is the diesel tank fuel percentage; is the battery pack power percentage; accurately estimate the hidden state variables in a narrow lane nonlinear environment.

[0008] Further, the S3 specifically comprises: S31: setting the prediction time domain step, the sampling time second; according to the diesel engine torque characteristic curve, the electric pump response characteristic and the pressure-flow relationship of the hydraulic mechanism I and the hydraulic mechanism II, a discrete time prediction model is established ; wherein is the prediction step number; is the sampling interval; is the state transition matrix; To control the input matrix; To the state vector of the first step; To the control vector of the first step; S32: Constructing a quadratic objective function ; Wherein is the objective function value; is the total energy consumption of the first step, in kilowatt-hour; is the reference energy consumption; is the vehicle inclination angle of the first step; is the wheel slip rate of the first step; is the power change rate; , , is the state weight coefficient; is the control weight coefficient; S33: Establishing equality constraints and inequality constraints , , ; Wherein is the power of the diesel engine; is the power of the electric pump; is the total power demand; is the battery capacity; kW is the unit of kilowatt; S34: Using the sequential quadratic programming method to solve the optimization problem to obtain the power distribution sequence of the future step ; Taking the optimization result of the first step , as the power distribution instruction of the current time; Wherein the superscript represents the optimal value; The first-level distribution takes 80% as the basic power, the second-level distribution takes as the compensation power, and the third-level distribution takes as the peak buffer power; Under the condition of meeting the operation demand, the energy consumption is optimized and the stability is best.

[0009] Further, the S4 specifically comprises: S41: reading the actual output power of the diesel engine and the actual output power of the electric pump every second; Calculating the power tracking error ; When , adjusting the throttle opening of the diesel engine or the speed of the electric pump through the PID controller in the control system; wherein is the feedback period; is the power error; Denotes absolute value operator; S42: judge uncertainty factor Whether greater than threshold value ; if , switch to conservative mode: set the target power ratio of the electric pump to 60%, and the target power ratio of the diesel engine to 40%; the diesel engine runs at a reduced speed in the conservative mode, reducing noise and emissions in the narrow lane environment; wherein Is the uncertainty threshold value; S43: judge the vehicle inclination Or Whether ; if yes, send a restriction signal to the control system, prohibit the arm extension action of the main telescopic arm and the amplitude action of the hydraulic mechanism I through the control box, and only allow the arm retraction and arm lowering actions; judge the working bucket load Whether And whether the working amplitude is Meters; if both conditions are met, trigger torque protection, prohibit all extension actions of the hydraulic mechanism I and the hydraulic mechanism II through the control system; S44: detect the vehicle brake signal; when detecting braking, turn off the pumping function of the electric pump through the control system, and switch the electric pump to the power generation mode; use the hydraulic energy in the hydraulic circuit to drive the electric pump to generate electricity, and the generated electric energy is used to charge the battery pack; record the recovered electric quantity ; wherein Is the recovered electric quantity, unit: ampere-hour; Is the charging current; Denotes integral operator; Is the time differential; ensure the safety of narrow lane operation and improve energy utilization efficiency.

[0010] The beneficial effects achieved by the present application are: First, the present application designs a logic controller to handle the uncertainty of narrow lane terrain, which quantifies the uncertainty factor by fuzzy processing the vehicle inclination, working bucket load and wheel speed data, converts the complex terrain features that are difficult to describe with precise mathematical models into controllable quantitative parameters, so that the system can automatically adjust the power distribution strategy according to the complexity of the terrain, significantly improving the terrain adaptability and operation reliability of the hybrid operation platform in narrow and complex environments such as alleys.

[0011] Secondly, the particle filtering algorithm is used to replace the traditional Kalman filtering to estimate the hidden state, and through the search of one hundred particles in the nonlinear state space, the key state variables such as wheel slip rate and energy reserve which are difficult to directly measure are accurately estimated, the strong nonlinear and non-Gaussian characteristics of vehicle dynamics in narrow lane environment are effectively handled without the restriction of linear and Gaussian assumption, the problem of insufficient estimation accuracy of traditional filtering method in complex environment is overcome, and accurate and reliable state information basis is provided for subsequent power distribution decision, thereby the precision and robustness of the whole control system are improved.

[0012] Thirdly, the model predictive control technology is applied to multi-stage power optimization distribution, a prediction time domain is set, a quadratic objective function including energy consumption term, inclination term and slip rate term is constructed, under the premise of meeting the power balance constraint and battery power constraint, the sequence quadratic programming method is used to solve the optimization problem, the coordinated control of primary diesel power distribution, secondary electric power distribution and tertiary hydraulic peak buffer distribution is realized, the future state evolution is predicted, the global optimal energy management is realized through rolling optimization, compared with the traditional rule-based switching strategy, the local suboptimal solution is avoided, the system energy consumption is significantly reduced, and the stability of power output is improved.

[0013] Fourthly, the real-time state acquisition module, fuzzy preprocessing unit, particle filtering estimation unit, MPC predictive control unit and adaptive control unit are designed, and a complete intelligent control system architecture is formed. The system adopts the hybrid configuration of diesel engine and electric pump, through adaptive feedback and switching mechanism, the normal mode and the conservative mode are automatically switched according to the uncertainty factor, the electric power proportion is increased and the diesel engine speed is reduced in the conservative mode, and the brake energy recovery function is realized at the same time, the energy utilization efficiency is improved, the noise and emission pollution are reduced, and the applicability in sensitive environmental areas such as narrow lane is enhanced.

[0014] Fifthly, the fuzzy logic, particle filtering and model predictive control are cooperated to realize the intelligent power distribution of the hybrid working platform in the narrow lane complex environment, the power combination of the diesel engine and the electric pump is automatically adjusted according to the real-time working condition, the comprehensive target of optimal energy consumption and best stability is realized under the premise of ensuring the working demand. The overall scheme is especially suitable for power distribution line operation in narrow area, solves the technical problems of high energy consumption, poor stability and insufficient environmental adaptability of the traditional single power working vehicle in narrow lane environment, and provides an efficient and energy-saving technical solution for the maintenance operation of urban power distribution system. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 It is the structure diagram of the hybrid working platform suitable for narrow lane operation of the application Figure 1 ; Figure 2is a structure diagram of a hybrid working platform suitable for narrow lane operation of the present application Figure 2 ; Figure 3 is a power distribution system framework diagram of a hybrid working platform suitable for narrow lane operation of the present application Figure 4 is a power distribution method flow chart of a hybrid working platform suitable for narrow lane operation of the present application Figure 5 is a comparison curve diagram of diesel power, electric power and total power changing with time of the present application embodiment 2 and each comparative example Figure 6 is a comparison curve diagram of wheel slip rate change of the present application embodiment 2 and each comparative example in the uphill section for 180 seconds Figure 7 is a comparison bar chart of comprehensive energy consumption amount of the present application embodiment 2 and each comparative example for completing the same working task

[0016] Figure mark: 1, walking platform; 2, working device; 3, control system; 11, wheeled four-wheel drive chassis; 111, steering mechanism; 12, diesel engine; 13, diesel tank; 14, electric pump; 15, battery pack; 21, rotary table; 22, main telescopic arm; 23, hydraulic mechanism I; 24, insulated arm; 25, hydraulic mechanism II; 26, slewing mechanism; 27, working bucket; 31, control box; 32, remote controller. DETAILED DESCRIPTION

[0017] The technical solutions in the present application will be described clearly and completely in combination with the drawings in the present application, and in addition, the forms of each structure described in the following embodiments are only examples, and the present application is not limited to each structure described in the following embodiments, and all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0018] Reference Figures 1-4The application provides a hybrid working platform power distribution method for narrow lane operation. The method realizes power optimization distribution of a diesel engine and an electric pump based on real-time collection of vehicle state information and fuzzy logic, particle filtering and model predictive control technology. The overall process of the method includes four main steps of S1 real-time state collection and fuzzy preprocessing, S2 particle filtering hidden state estimation, S3 multi-stage prediction distribution based on MPC and S4 adaptive feedback and switching. The four steps are executed in time sequence to form a closed-loop control system. Each control cycle starts from state collection, quantifies terrain uncertainty through fuzzy preprocessing, estimates hidden state variables that are difficult to directly measure through particle filtering, optimizes power distribution strategy in the future time domain based on a prediction model, and finally performs adaptive adjustment and safety protection according to the actual execution. The whole cycle takes about 100-150 milliseconds, which can meet the real-time control requirements of narrow lane operation.

[0019] In step S1, the inclination and acceleration data of the vehicle are collected at a high frequency by an IMU sensor, i.e. an inertial measurement unit, and the working bucket load and wheel speed information are obtained by a load sensor and a wheel speed sensor. Then, the continuous physical quantities are converted into discrete fuzzy sets by a fuzzy logic controller, i.e. FLC, and finally an uncertainty factor between 0 and 1 is output according to the pre-established fuzzy rule base for quantifying the complexity of the narrow lane terrain. The core idea of this step is to convert the characteristics of the narrow lane terrain, which are difficult to describe with an accurate mathematical model, into quantifiable control parameters through fuzzy processing, providing terrain adaptability basis for subsequent power distribution decisions.

[0020] Specifically, step S11 collects the front and rear inclination angles of the vehicle and the left and right inclination angles at a sampling frequency of 100Hz by the IMU sensor. The selection of the sampling frequency is based on the dynamic response characteristics of the vehicle, and 100Hz can capture the rapid inclination changes caused by the bumps of the narrow lane road without generating excessive redundant data. The unit of the inclination angle is degree. When the absolute value of the inclination angle is or , the system determines that the vehicle is in a dangerous inclination state and sets the inclination angle alarm flag to 1. This flag will trigger the protection limiting action in the subsequent S4 step. The determination of the inclination angle threshold is based on vehicle stability analysis. When the wheeled four-wheel drive chassis can still keep four wheels on the ground and the center of gravity projection is within the support polygon at this inclination angle, exceeding this threshold value will have the risk of overturning.

[0021] Step S12 collects the current load weight of the working bucket , the sampling frequency is lower than IMU sensors because load changes relatively slowly, 50Hz is enough to track load changes caused by the operator moving the bucket or carrying materials. The unit of load is kilogram kg, and the measurement range is 0-150kg corresponding to the rated load of the bucket. At the same time, the wheel speed sensors installed in the hubs of the four wheels collect the front left wheel speed , front right wheel speed , rear left wheel speed and rear right wheel speed , the unit of wheel speed is revolutions per minute rpm. The four independent wheel speed signals are not only used to calculate the average speed of the vehicle, but also can reflect the slip state of each wheel. When a certain wheel slips on wet or soft ground, its speed will be significantly higher than that of other wheels, and this information is of great value for subsequent particle filter estimation of wheel slip rate .

[0022] Step S13 defines a triangular membership function according to the collected inclination and load data, which maps continuous physical quantities to discrete fuzzy sets. For inclination , when , it is completely affiliated with the "low" fuzzy set, indicating a flat terrain, with a membership degree of 1; when , it is partially affiliated with the "low" and "medium" fuzzy sets, with a membership degree changing linearly, for example, at , the membership degrees of "low" and "medium" are both 0.5; when , it is completely affiliated with the "high" fuzzy set, indicating a steep terrain. The advantage of the triangular membership function is simple calculation and clear physical meaning, but it is not differentiable at the boundary, which has little effect in practical application. Similarly, for load , when , it is affiliated with the "light" fuzzy set, when , it is affiliated with the "medium" fuzzy set, and when , it is affiliated with the "heavy" fuzzy set. For average speed , first calculate the four-wheel average speed by the formula , when , it is affiliated with the "slow" fuzzy set, when , it is affiliated with the "medium" fuzzy set, and when , it is affiliated with the "fast" fuzzy set. The parameters of these membership functions are set based on statistical analysis of a large amount of field test data, which can well cover the typical working conditions of narrow lane operation.

[0023] Step S14 establishes a fuzzy rule base containing 27 rules, the number of rules is equal to the Cartesian product of three inputs each with three fuzzy sets . Each rule uses the IF-THEN format, for example, "IF inclination is high AND load is heavy AND speed is slow THEN uncertainty factor “High”, the output value corresponding to the rule is 0.9. All 27 rules cover all possible combinations of the three input variables of inclination, load and speed, ensuring that at least one rule is activated for any input state.

[0024] The design of fuzzy rules follows the following principles: first, the inclination has the greatest weight on uncertainty because it is directly related to the stability and rollover risk of the vehicle, and is the most critical safety factor in narrow lane working environment; second, the load has the second greatest weight on uncertainty because it affects the vehicle's center of gravity position and power demand, and heavy load increases the risk of overturning and increases the requirements on the power system; third, the speed has a relatively small weight on uncertainty, but fast driving will shorten the response time and increase the difficulty of dynamic control. In terms of specific numerical settings, when all three input variables are at the lowest level, i.e. low inclination, light load and slow speed, the uncertainty factor of the corresponding rule R1 is set to 0.1, indicating that the terrain environment is simple and certain; when all three input variables are at the highest level, i.e. high inclination, heavy load and fast speed, the uncertainty factor of the corresponding rule R27 is set to 0.9, indicating that the terrain environment is extremely complex and uncertain; for intermediate working conditions, the uncertainty factor is continuously distributed between 0.1 and 0.9 according to the combination of each input variable, reflecting the gradual change characteristics of terrain complexity. The output uncertainty factor α of the fuzzy rule base plays a key role in subsequent control. When the uncertainty factor α is greater than the threshold value 0.8, the system automatically switches to the conservative mode, increasing the electric power ratio from the normal 20% to 60%, and the diesel power ratio correspondingly decreases from 80% to 40%. In the conservative mode, the advantage of fast response speed of the electric pump is fully utilized, which can quickly respond to sudden load changes in complex terrain, while the diesel engine runs at a reduced speed to reduce noise and emissions in narrow lane environment, reflecting the advantages of the invention in safety and environmental protection.

[0025] In a preferred embodiment, the fuzzy rule base comprises 27 rules for outputting an uncertainty factor a according to the fuzzy set combination of three input variables of vehicle inclination, working bucket load and wheel speed. The uncertainty factor a is used to quantify the complexity and uncertainty level of the terrain environment in the narrow lane working environment, and its value ranges from 0.1 to 0.9, and the larger the value, the more complex the terrain environment and the higher the uncertainty. The fuzzy set division of the three input variables is as follows: the vehicle inclination is divided into three fuzzy sets of low, medium and high, wherein the low level corresponds to an inclination less than 3 degrees, the medium level corresponds to an inclination greater than or equal to 3 degrees and less than 6 degrees, and the high level corresponds to an inclination greater than or equal to 6 degrees; the working bucket load is divided into three fuzzy sets of light, medium and heavy, wherein the light level corresponds to a load less than 50 kg, the medium level corresponds to a load greater than or equal to 50 kg and less than 100 kg, and the heavy level corresponds to a load greater than or equal to 100 kg; the wheel speed is divided into three fuzzy sets of slow, medium and fast, wherein the slow level corresponds to an average speed less than 500 revolutions per minute, the medium level corresponds to an average speed greater than or equal to 500 revolutions per minute and less than 1000 revolutions per minute, and the fast level corresponds to an average speed greater than or equal to 1000 revolutions per minute. Specifically as follows: Rule R1: When the vehicle inclination is at low level, the load in working bucket is at light level and the wheel speed is at slow level, it indicates that the vehicle is driving on flat road at low speed with empty load, the terrain environment is simple, the system sets the uncertainty factor a as 0.1. Rule R2: When the vehicle inclination is at low level, the load in working bucket is at light level and the wheel speed is at medium level, it indicates that the vehicle is driving on flat road at medium speed with empty load, the terrain uncertainty increases slightly, the system sets the uncertainty factor a as 0.2. Rule R3: When the vehicle inclination is at low level, the load in working bucket is at light level and the wheel speed is at fast level, it indicates that the vehicle is driving on flat road at fast speed with empty load, the dynamic response demand increases due to the fast speed, the system sets the uncertainty factor a as 0.3. Rule R4: When the vehicle inclination is at low level, the load in working bucket is at medium level and the wheel speed is at slow level, it indicates that the vehicle is driving on flat road at low speed with load, the uncertainty increases slightly due to the load, the system sets the uncertainty factor a as 0.2. Rule R5: When the vehicle inclination is at low level, the load in working bucket is at medium level and the wheel speed is at medium level, it indicates that the vehicle is driving on flat road at medium speed with load, it belongs to the normal working condition, the system sets the uncertainty factor a as 0.3. Rule R6: When the vehicle inclination is at low level, the load in working bucket is at medium level and the wheel speed is at fast level, it indicates that the vehicle is driving on flat road at fast speed with load, the combination of load and speed increases the fluctuation of power demand, the system sets the uncertainty factor a as 0.4. Rule R7: When the vehicle inclination is at low level, the load in working bucket is at heavy level and the wheel speed is at slow level, it indicates that the vehicle is driving on flat road at low speed with heavy load, the heavy load increases the power demand but the slow speed makes the control relatively easy, the system sets the uncertainty factor a as 0.3. Rule R8: When the vehicle inclination is at low level, the load in working bucket is at heavy level and the wheel speed is at medium level, it indicates that the vehicle is driving on flat road at medium speed with heavy load, the combination of heavy load and medium speed increases the burden of the system, the system sets the uncertainty factor a as 0.4. Rule R9: When the vehicle inclination is at low level, the load in working bucket is at heavy level and the wheel speed is at fast level, it indicates that the vehicle is driving on flat road at fast speed with heavy load, the combination of heavy load and fast speed requires high demand for the power system, the system sets the uncertainty factor a as 0.5.

[0026] When the vehicle is in the medium level of the inclination angle, the medium level of the load of the working bucket and the medium level of the wheel speed, it indicates that the vehicle is in the medium speed load driving on the slope, which belongs to the medium complex working condition, the system sets the uncertainty factor a as 0.5. Rule R15: When the vehicle is in the medium level of the inclination angle, the medium level of the load of the working bucket and the fast level of the wheel speed, it indicates that the vehicle is in the fast load driving on the slope, all the three factors are in the medium level and above, the system sets the uncertainty factor a as 0.6. Rule R16: When the vehicle is in the medium level of the inclination angle, the heavy level of the load of the working bucket and the slow level of the wheel speed, it indicates that the vehicle is in the slow heavy load driving on the slope, the heavy load and the inclination angle combination form a big challenge to the power system, the system sets the uncertainty factor a as 0.5. Rule R17: When the vehicle is in the medium level of the inclination angle, the heavy level of the load of the working bucket and the medium level of the wheel speed, it indicates that the vehicle is in the medium speed heavy load driving on the slope, the adverse combination of the terrain and the load increases the uncertainty, the system sets the uncertainty factor a as 0.6. Rule R18: When the vehicle is in the medium level of the inclination angle, the heavy level of the load of the working bucket and the fast level of the wheel speed, it indicates that the vehicle is in the fast heavy load driving on the slope, which belongs to the high risk complex working condition, the system sets the uncertainty factor a as 0.7.

[0027] Three, 9 rules under high inclination conditions; Rule R19: When the vehicle inclination is at high level, the working bucket load is at light level and the wheel speed is at slow level, it indicates that the vehicle is driving on steep slope at low speed and empty load, the high inclination leads to high uncertainty of the terrain environment, the system sets the uncertainty factor a as 0.5. Rule R20: When the vehicle inclination is at high level, the working bucket load is at light level and the wheel speed is at medium level, it indicates that the vehicle is driving on steep slope at medium speed and empty load, the combination of high inclination and medium speed increases the risk of rollover, the system sets the uncertainty factor a as 0.6. Rule R21: When the vehicle inclination is at high level, the working bucket load is at light level and the wheel speed is at fast level, it indicates that the vehicle is driving on steep slope at fast speed and empty load, fast driving under high inclination belongs to dangerous working condition, the system sets the uncertainty factor a as 0.7. Rule R22: When the vehicle inclination is at high level, the working bucket load is at medium level and the wheel speed is at slow level, it indicates that the vehicle is driving on steep slope at low speed and load, the combination of high inclination and load constitutes a severe test to stability, the system sets the uncertainty factor a as 0.6. Rule R23: When the vehicle inclination is at high level, the working bucket load is at medium level and the wheel speed is at medium level, it indicates that the vehicle is driving on steep slope at medium speed and load, the terrain is complex and the load is moderate, the system sets the uncertainty factor a as 0.7. Rule R24: When the vehicle inclination is at high level, the working bucket load is at medium level and the wheel speed is at fast level, it indicates that the vehicle is driving on steep slope at fast speed and load, the combination of high inclination, medium load and fast speed belongs to extreme working condition, the system sets the uncertainty factor a as 0.8. Rule R25: When the vehicle inclination is at high level, the working bucket load is at heavy level and the wheel speed is at slow level, it indicates that the vehicle is driving on steep slope at low speed and heavy load, the combination of high inclination and heavy load constitutes a great challenge to vehicle stability and power system, the system sets the uncertainty factor a as 0.7. Rule R26: When the vehicle inclination is at high level, the working bucket load is at heavy level and the wheel speed is at medium level, it indicates that the vehicle is driving on steep slope at medium speed and heavy load, all the three input variables are at disadvantageous level, the terrain environment is extremely complex, the system sets the uncertainty factor a as 0.8. Rule R27: When the vehicle inclination is at high level, the working bucket load is at heavy level and the wheel speed is at fast level, it indicates that the vehicle is driving on steep slope at fast speed and heavy load, all the three input variables reach the highest level, the terrain environment is extremely uncertain and dangerous, the system sets the uncertainty factor a as 0.9, at this time the system will switch to conservative mode to increase the electric power ratio to the maximum to improve the response speed and safety.

[0028] The fuzzy reasoning adopts the Mamdani method, which first calculates the activation degree of each rule , the activation degree is equal to the minimum value of the input membership degrees in the antecedent of the rule, which embodies the logical AND operation. Then defuzzification is carried out by the center of gravity method, the calculation formula is , wherein represents the summation of all 27 rules. For the first The activation level of the rule; For the first The output value corresponding to each rule. The physical meaning of the centroid method is to take a weighted average of the outputs of all activated rules, with the weight being the activation degree, and the output result is... It is a continuous value between 0 and 1. A value close to 0 indicates high terrain certainty and a simple environment. A value close to 1 indicates high uncertainty and complex environment. This fuzzification process transforms the difficult-to-describe concept of narrow alleyways and complex terrain into quantitative control parameters, providing terrain-adaptive capabilities for subsequent power allocation.

[0029] In step S2, the particle filter algorithm, i.e., PF, is used to estimate the wheel slip ratio. and energy surplus Two key hidden state variables. Particle filtering is a nonlinear filtering algorithm based on the Monte Carlo method, particularly suitable for handling the nonlinearity and non-Gaussianity of vehicle dynamics in narrow alley environments. Traditional Kalman filters (KF) assume a linear system model and that noise follows a Gaussian distribution. These assumptions often fail in complex terrain like narrow alleys. Particle filtering, however, approximates the posterior probability distribution of the state using a set of weighted particles, unrestricted by linear and Gaussian assumptions. Wheel slip ratio. Defined as the ratio of wheel slip distance to wheel rolling distance, it reflects the adhesion between the tire and the ground, particularly on wet or soft surfaces. A larger value indicates severe slippage. Energy reserve. It equals the sum of the diesel tank fuel percentage and the battery charge percentage, reflecting the system's energy reserve status. When the values ​​are low, energy-saving strategies need to be adopted. These two state variables are difficult to measure directly but are crucial to power allocation decisions, so they need to be estimated using filtering algorithms.

[0030] Step S21 Initialization The choice of the number of particles requires a trade-off between estimation accuracy and computational complexity. Too few particles will lead to inaccurate state estimation, while too many particles will increase the computational burden. 100 particles provide sufficient estimation accuracy while ensuring real-time performance in this application. The initial state vector of each particle is... The superscript 0 indicates the initial time, and the subscript... Indicates the first There are 10 particles, and the superscript T indicates vector transpose. The initial wheel slip ratio can be set according to the current ground conditions, for example, 0.05 on dry hard surfaces and 0.15 on wet and slippery surfaces; is the initial diesel engine speed, in rpm, which can be set to the idle speed 800 rpm; is the initial electric pump power, in kW, which can be set to 0 to indicate that the electric pump is not started; is the initial battery power percentage, in the range of 0 to 1, which can be set according to the reading of the battery management system (BMS). The initial weight of each particle is set to , indicating that all particles have the same importance at the initial moment. The initialization of particles adopts the method of adding random perturbations near the initial state, and the perturbation amplitude is determined according to the uncertainty of the state variable, which can disperse the particles in the state space and improve the search ability for the true state.

[0031] Step S22 performs state prediction for each particle, and the prediction formula is , where is the predicted state of the th particle at the th moment; is a nonlinear state transition function describing how the system state evolves over time; is the state of the th particle at the th moment; is the control input vector at the th moment, including the throttle opening of the diesel engine and the start-stop signal of the electric pump; is the process noise, representing model uncertainty and external disturbance, which is assumed to follow a multi-dimensional Gaussian distribution with mean 0 and covariance matrix The specific form of the state transition function is established based on the vehicle dynamics model, for example, the evolution of the wheel slip ratio can be described by the tire-ground interaction model, the evolution of the diesel engine speed can be described by the engine dynamics equation, and the evolution of the electric pump power and the battery power can be described by the electrical system model. Due to the complexity of the narrow lane environment, these models often exhibit strong nonlinear characteristics, such as the nonlinear relationship between tire friction and slip ratio, and particle filtering can well handle this nonlinearity. The elements of the process noise covariance matrix need to be determined according to system identification experiments or expert experience, and the diagonal elements reflect the uncertainty of each state variable, and the non-diagonal elements reflect the correlation between states.

[0032] Step S23 first acquires the sensor observation value , where is the average wheel speed observed at the th moment, is the average wheel speed observed at the Battery current observed at time t, positive for discharging, negative for charging, in amperes, A. Then compute the likelihood of each particle where denotes the natural exponential function, i.e. denotes the Euclidean norm of a vector, i.e. the square root of the sum of the squares of the vector components; is the observation function, mapping the state space to the observation space, e.g. the average wheel speed can be computed from the vehicle speed and the slip ratio in the state vector, the battery current can be computed from the electric motor power and the battery internal resistance; is the observation noise covariance matrix, reflecting the uncertainty of the sensor measurements. The computation of the likelihood embodies the idea of Bayesian inference: if the state of a particle is close to the actual observation, the likelihood of this particle is high, indicating that this particle is more likely to represent the true state. Next, the weight of each particle is updated, the update formula is i.e. the new weight is equal to the old weight multiplied by the likelihood, this step realizes the Bayesian update. Then the weights are normalized where is the sum index from 1 to summing over all particles, normalization ensures that the sum of all weights is 1, so that the weights can be interpreted as probabilities. After normalization, particles with larger weights indicate that their states are closer to the true state, and have greater influence in subsequent state estimation.

[0033] Step S24 first computes the effective number of particles , the effective number of particles reflects the concentration of the particle distribution, if all particle weights are equal , if there is only one particle with weight 1 and the rest are 0 . When , i.e. the effective number of particles is less than half of the total number of particles, it means that most of the particles have very small weights and contribute little to the state estimation, there is a phenomenon of particle degeneracy, at this time the resampling operation needs to be performed. The basic idea of resampling is to resample particles according to their weights, keeping high-weight particles and eliminating low-weight particles, so as to improve the effectiveness of particles. Specifically, the systematic resampling algorithm is used, which first generates random numbers uniformly distributed in the interval [0, 1), then selects particles according to the cumulative weight distribution, particles with large weights will be selected many times and particles with small weights may be eliminated. After resampling, all weights are reset to , so that the particles have the same importance again. Finally, the wheel slip ratio estimate is calculated, i.e. the weighted average of the slip ratio states of all particles, with the weight being the normalized particle weight. Similarly, the energy margin estimate is calculated, where ​Diesel tank fuel percentage, Battery pack state of charge. This step realizes the conversion from particle set to single state estimation value, providing accurate hidden state information for subsequent power allocation decisions. The particle filter method can effectively handle nonlinear and non-Gaussian characteristics in narrow lane environments by introducing a particle swarm to represent the state distribution. Compared with traditional Kalman filtering, it has higher estimation accuracy, especially when the vehicle encounters sudden terrain changes such as suddenly driving onto a slope. Particle filtering can quickly adjust state estimation without significant estimation bias.

[0034] In step S3, a model predictive control algorithm, MPC, is used for multi-level power allocation. MPC is an advanced control method based on system prediction model, which solves an optimization problem within a future time window according to the current state of the system and the prediction model at each control time, and obtains the optimal control sequence. Then only the first control amount of the sequence is executed, and the above process is repeated at the next control time to form rolling optimization. The advantage of MPC is that it can explicitly handle system constraints such as power upper and lower limits, battery charge limits, etc., and can predict future state evolution to make forward-looking control decisions, which is of great significance to improve the energy efficiency and operational stability of hybrid systems.

[0035] Step S31 first sets the prediction horizon The selection of the prediction horizon needs to balance between prediction performance and computational complexity. A too short prediction horizon will result in a lack of foresight in the control strategy, while a too long prediction horizon will significantly increase the optimization calculation. A 10-step prediction can provide sufficient foresight in this application, with a preferred range of 5 to 20 steps. The sampling time is 0.5 seconds, indicating that the control amount is updated every 0.5 seconds. The selection of the sampling time needs to be faster than the response speed of the system's dominant dynamics, and 0.5 seconds can track the power changes of the hybrid system without producing excessive computational burden, with a preferred range of 0.2 seconds to 1 second. Therefore, the prediction horizon covers a time window of 5 seconds in the future. Then a discrete-time prediction model is established according to the torque characteristic curve of the diesel engine, the response characteristics of the electric pump, and the pressure-flow relationship of hydraulic mechanism I and hydraulic mechanism II , where is the state vector of the first step, containing state variables such as vehicle speed, inclination angle, and slip rate; is the control vector of the first step, containing control inputs such as diesel throttle opening and electric pump speed; The input matrix is for control. The model is obtained by linearizing the nonlinear system model around the operating point or by system identification method, although it is a linear model, it can well approximate the system dynamics near the operating point. The accuracy of the prediction model directly affects the control performance of MPC, therefore, the model parameters need to be updated regularly according to the actual operation data to adapt to the changes in system characteristics.

[0036] Step S32 constructs a quadratic objective function, which quantifies the control objectives including minimizing energy consumption, reducing the pitch angle fluctuation and reducing the wheel slip rate, while penalizing the sharp change of power. The mathematical expression of the objective function is:

[0037] ; Where is the value of the objective function; the first summation term is the sum of the first steps in the prediction horizon, the second summation term is the sum of the first steps in the prediction horizon, because the last step has no subsequent steps, it cannot calculate the power change rate; is the total energy consumption of the step, the unit is kilowatt-hour kWh, which is obtained by integrating the instantaneous power; is the reference energy consumption, which is set according to historical data or expert experience, representing the ideal energy consumption level; is the pitch angle of the vehicle at the step, the unit is degree; is the wheel slip rate at the step, dimensionless; is the power change rate; , , is the state weight coefficient, used to balance the importance of different control objectives, the preferred value is , , These weights can be adjusted according to the actual application requirements, for example, if more attention is paid to energy consumption, the value of can be increased, if more attention is paid to stability, the value of and can be increased; is the control weight coefficient, the preferred value is 0.1, this coefficient is used to smooth the change of control quantity to avoid frequent switching leading to wear of the actuator. The reason for using the quadratic form of the objective function is that the quadratic function has convexity, its optimization problem has a unique optimal solution and can be solved efficiently, and the quadratic term can effectively penalize the state deviating from the target value.

[0038] Step S33 establishes equality constraints and inequality constraints to reflect the physical limitations of the system. The equality constraint represents that the sum of the diesel engine power and the electric pump power must be equal to the total power demand, where Power for diesel engine, Power for electric pump, The total power demand is determined by the hydraulic load of the work device. This constraint ensures power balance, avoiding slow operation of the work device due to insufficient power or energy waste due to excessive power. Inequality constraints include represents the upper and lower limits of diesel engine power, where 35.9kW is the rated power of the Yanmar diesel engine, and the lower limit of 0 indicates that the diesel engine can be turned off; represents the upper and lower limits of electric pump power, where 15kW is the rated power of the electric pump; represents the upper and lower limits of the battery power percentage, with the lower limit of 0.2 to protect the battery from over-discharge and the upper limit of 0.95 to avoid overcharging, based on the manufacturer's usage recommendations. By explicitly handling these constraints, the MPC can ensure that the optimization result is within the physically feasible range, avoiding violation of system limits that can cause equipment damage or performance degradation.

[0039] Step S34 solves the constrained optimization problem using the Sequential Quadratic Programming (SQP) method. SQP is an iterative optimization algorithm that transforms the nonlinear optimization problem into a series of quadratic programming subproblems through quadratic approximation. Each iteration solves a QP subproblem to obtain a search direction and step size, and then updates the variables until convergence. SQP algorithm is suitable for solving constrained nonlinear optimization problems, with fast convergence speed and good numerical stability. The solver outputs the optimal power allocation sequence for the future steps where the superscript "*" represents the optimal value. Since the MPC adopts a rolling optimization strategy, only the first step of the optimization result and is sent to the actuator as the current power allocation instruction, and the power allocation of future steps will be re-optimized at the next control time. This approach can continuously correct the control strategy based on real-time feedback of system state, improving robustness. Multi-level allocation reflects that the first level allocation takes 80% of the total power as the base power, which undertakes the main power output task, taking advantage of the large torque of the diesel engine; the second level allocation takes as compensation power, which fine-tunes the diesel power, taking advantage of the fast response and high control accuracy of the electric pump; the third level allocation takes As a peak buffer power source, it absorbs the instantaneous high-load impact of the hydraulic system, such as the peak power demand during rapid boom extension. This hierarchical coordination leverages the advantages of each power source while avoiding overload operation of a single power source, thus improving the system's energy efficiency and lifespan. The MPC predictive control method, by predicting future state evolution and optimizing the control sequence, can achieve global optimization rather than local optimization for the hybrid system. Especially in narrow-lane operations, when the vehicle anticipates an incline ahead, MPC can adjust power distribution in advance to increase diesel power reserves, preventing insufficient power during uphill climbing.

[0040] In step S4, adaptive feedback and switching control are executed, adjusting the power allocation strategy according to the actual execution situation, and implementing safety protection and energy recovery. This step is the execution and protection layer of the entire control method, ensuring that control commands can be executed accurately and that timely intervention is possible in abnormal situations.

[0041] Step S41 every Read the actual output power of the diesel engine in seconds and the actual output power of the electric pump Feedback cycle The preferred range is 0.05 seconds to 0.2 seconds; 0.1 seconds allows for rapid tracking of power changes without placing an excessive burden on the communication system. Then, the power tracking error is calculated. The minus sign indicates arithmetic subtraction, and parentheses indicate that the addition within the parentheses should be calculated first, followed by the subtraction. A positive value indicates that the actual power is less than the demand, indicating a power shortage. A negative value indicates that the actual power exceeds the demand, resulting in power surplus. When the absolute value of the power error... If the error exceeds 0.5 kW, the throttle opening of the diesel engine or the speed of the electric pump is adjusted by the PID controller (Proportional-Integral-Derivative) in the control system. The PID controller is a classic feedback controller. It uses a proportional term to quickly respond to errors, an integral term to eliminate steady-state errors, and a derivative term to predict error trends. The weighted combination of these three terms enables precise power tracking. The 0.5 kW error threshold is set to consider control accuracy requirements and avoid control oscillations; this value is approximately 1% of the maximum power, with an optimal range of 0.2 kW to 1 kW. This feedback control loop can promptly correct power deviations caused by model errors, external disturbances, or actuator delays, ensuring that the actual power output meets operational requirements.

[0042] Step S42: Determine the uncertainty factor Is it greater than the threshold? The threshold is set based on safety considerations for narrow aisle operations. When the terrain is uncertain, such as wet road surface, steep slope or excessive load, the system switches to the conservative mode. In the conservative mode, the target power ratio of the electric pump is increased from 20% in the normal mode to 60%, and the target power ratio of the diesel engine is decreased from 80% to 40% accordingly. The reason for this switching strategy is that the electric pump has low noise, zero emissions and fast response when driven. Increasing the electric power ratio in uncertain environments can reduce the noise and pollution impact on the surrounding environment, and the fast response characteristics of the electric pump help to respond to sudden load changes. The diesel engine runs at a lower speed in the conservative mode, from a high speed of 2400 rpm to a medium speed of about 1600 rpm. Reducing the speed can significantly reduce noise and emissions, although the output power has decreased, but the overall power supply can still meet the demand because the electric pump bears more power. Threshold The preferred range is 0.7 to 0.9, which can be adjusted according to the noise limit and emission requirements of the actual working environment. This adaptive mode switching embodies the intelligent characteristics of the system, which can automatically adjust the power distribution strategy according to the complexity of the terrain, and reduce the environmental impact as much as possible under the premise of ensuring the working ability.

[0043] Step S43 implements double protection judgment. First, it is judged whether the vehicle inclination or is If the condition is met, a limit signal is sent to the control system, and the control box is prohibited from stretching the main telescopic arm and the amplitude action of the hydraulic mechanism I, and only the arm and the arm action are allowed. The direction of movement is allowed to move towards the body. The purpose of this protection measure is to prevent the extension of the arm in the inclined state from causing the vehicle center of gravity to shift and increase the risk of overturning, while the arm and the arm action will move the center of gravity to the center of the vehicle. Therefore, it is allowed to increase stability. Inclination protection is based on vehicle statics analysis. When the inclination reaches and the arm is in the maximum extension position, the whole vehicle is close to the stability limit, and further extension of the arm may cause a rollover accident. Then it is judged whether the working bucket load is and whether the working amplitude is If both conditions are met at the same time, the torque protection is triggered, and all extension actions of hydraulic mechanism I and hydraulic mechanism II are prohibited by the control system. The principle of torque protection is based on the torque balance of the boom system. When the load reaches the rated value of 150 kg and the working range reaches the maximum value of 5.6 meters, the overturning torque generated by the boom on the turntable reaches the maximum allowed value. At this time, if the boom continues to extend or the load increases, the overturning torque will exceed the stable torque, causing the vehicle to tilt forward or even overturn. Torque protection evaluates the overturning risk by calculating the product of load and range in real time. This protection method is more scientific than simply limiting load or range because the overturning torque depends on the product of load and force arm. These two protection measures constitute a complete safety protection system, monitoring vehicle stability from two dimensions of inclination and torque, ensuring the safety of narrow lane operation, and preventing accidents even if the operator makes a mistake.

[0044] Step S44 realizes the braking energy recovery function. First, the vehicle braking signal is detected, which comes from the brake pedal switch or the brake button of the remote controller. When braking is detected, the pumping function of the electric pump is turned off by the control system, i.e., the motor-driven pump stops supplying oil to the hydraulic system, and the electric pump is switched to the power generation mode, i.e., the winding connection mode of the permanent magnet synchronous motor is switched from the motor mode to the generator mode. In the power generation mode, the hydraulic motor of the electric pump is driven in reverse by the hydraulic energy stored in the hydraulic circuit. Since the permanent magnet synchronous motor is reversible, the motor will generate an induced electromotive force when it reverses, which will output electric energy to the outside. This is the physical principle of braking energy recovery. The generated electric energy is charged to the battery pack through the power electronic module, i.e., the DC-DC converter and the charging controller. The charging process is monitored by the BMS battery management system to ensure that the charging current and voltage are within a safe range. The calculation formula for recording the recovered electric energy is wherein is the recovered electric energy, in ampere-hours Ah; is the charging current, in amperes A; is the integral operator with respect to time; is the time differential, in seconds s. The integral operation is realized by numerical integration methods such as the trapezoidal method or the Simpson method. The charging current is measured once every sampling period, and then the product of the current and time is accumulated to obtain the electric energy. Energy recovery not only improves the energy utilization efficiency of the system and reduces the dependence on external charging, but also prolongs the use time of the battery, which is of great significance in narrow lane continuous operation. According to the calculation, braking energy recovery can recover 30-50% of the braking energy, which is equivalent to extending the battery endurance time by 10-15%. This reflects the energy management advantage of the hybrid system, which converts the energy wasted during braking into electrical energy and stores it, achieving secondary utilization of energy.

[0045] This invention introduces fuzzy logic to handle the uncertainties of narrow alley terrain, transforming complex terrain features that are difficult to quantify into uncertainties that can be used for control decisions, thus providing terrain-adaptive capabilities for power allocation. Secondly, it employs a particle filter algorithm to estimate wheel slip ratio and energy margin, which are difficult to measure directly, overcoming the insufficient accuracy of traditional Kalman filtering in nonlinear, non-Gaussian environments, and providing accurate state information for power allocation decisions. Thirdly, it applies model predictive control for multi-level power optimization allocation, achieving globally optimal energy management by predicting future state evolution and rolling optimization control sequences. Compared to traditional rule-based switching strategies, this reduces energy consumption by 25% and improves stability by 30%. This method is particularly suitable for power distribution line operations in narrow and complex terrains such as the hutongs of Beijing's old city, significantly improving energy efficiency and environmental adaptability while ensuring operational capabilities, demonstrating high practical value and promising prospects for wider application.

[0046] Example 1, referring to Figures 1-4 This embodiment discloses a power distribution system for a hybrid work platform used in narrow alley operations. The system is installed on a traveling platform 1 and includes: a diesel engine 12, an electric pump 14, a battery pack 15, a control system 3, and a work device 2 connected thereto. The traveling platform 1 uses a wheeled four-wheel drive chassis 11, on which a steering mechanism 111, a diesel tank 13, and a hydraulic drive device are installed. The diesel engine 12 and the electric pump 14 are connected to the hydraulic lines of hydraulic mechanisms I 23 and II 25, respectively, providing power to the main telescopic boom 22, the insulated boom 24, and the work bucket 27. The control system 3 includes a control box 31 and a remote controller 32. The control box 31 is electrically connected to the diesel engine 12, the electric pump 14, the battery pack 15, and the sensor module for power distribution and feedback control. The control box 31 contains a status acquisition module, a fuzzy preprocessing unit, a particle filter estimation unit, a model predictive control (MPC) unit, and an adaptive control unit.

[0047] The state acquisition module is connected with the IMU sensor, the load sensor and the four wheel speed sensors installed on the walking platform 1, and is used for acquiring the vehicle inclination, the load and the wheel speed signals; the fuzzy preprocessing unit receives the above signals, and outputs the terrain uncertainty factor after fuzzy logic operation; the particle filtering estimation unit estimates the wheel slip rate and the energy margin according to the collected data; the model predictive control unit generates the diesel power and the electric power distribution instructions according to the energy consumption and the power constraint; the adaptive control unit receives the actual power feedback signal, adjusts the power ratio of the diesel engine and the electric pump when the uncertainty factor exceeds the threshold, and executes the mode switching, the inclination and load protection and the brake energy recovery. The output shaft of the diesel engine 12 drives the hydraulic mechanism I 23 through mechanical connection, and provides the speed signal feedback to the control system 3 through electrical connection; the electric pump 14 and the input end of the hydraulic mechanism II 25 are hydraulically connected, and are electrically connected with the output end of the battery pack 15, and are used for bidirectional energy conversion in the driving or energy recovery state.

[0048] The working device 2 is installed on the walking platform 1 through the turntable 21, the turntable 21 is connected with the slewing mechanism 26, one end of the main telescopic arm 22 is hinged to the turntable 21, and the other end is connected with the working bucket 27 through the insulating arm 24. The hydraulic mechanism I 23 is used for controlling the amplitude and telescopic action of the main telescopic arm 22, and the hydraulic mechanism II 25 is used for controlling the pitching action of the insulating arm 24 and the lifting action of the working bucket 27. The above hydraulic mechanisms are cooperatively provided with hydraulic power by the diesel engine 12 and the electric pump 14. The control box 31 is connected with the electromagnetic valve signals of the hydraulic mechanism I 23, the hydraulic mechanism II 25 and the slewing mechanism 26, and is used for controlling the hydraulic actuator to act according to the power distribution instructions output by the MPC unit, so as to realize the stable driving of the working arm. The remote controller 32 is wirelessly communicated with the control box 31, and is used for remotely controlling the telescopic, rotating and lifting of the working device 2.

[0049] When the vehicle is running or working in a narrow lane environment, the IMU sensor collects the vehicle inclination signal in real time, the load sensor collects the working bucket load, and the wheel speed sensor collects the four-wheel speed; the fuzzy preprocessing unit inputs the above signals into the fuzzy logic controller to output the uncertainty factor; the particle filter unit combines the factor and the energy state to estimate the optimal power distribution ratio of the diesel engine and the electric pump; the model predictive control unit solves the quadratic programming problem in the prediction time domain with several future sampling steps to output the diesel power and electric power distribution instructions at the current time. The adaptive control unit receives the actual output power feedback signals of the diesel engine 12 and the electric pump 14, calculates the power error, and adjusts the throttle opening and the electric pump speed through the PID controller; when the uncertainty factor exceeds the threshold value, the system automatically switches to the conservative mode to increase the electric power ratio to enhance the response speed; when the inclination or load is out of limit, the control box 31 outputs the protection signal to limit the action of the main telescopic arm 22 and the hydraulic mechanism I 23; when the brake signal is detected, the control system 3 controls the electric pump 14 to switch to the power generation mode, and the hydraulic energy is charged into the battery pack 15 through the power electronic module to realize energy recovery.

[0050] The entire power distribution system realizes the coordinated control of the diesel engine 12 and the electric pump 14 through the closed-loop structure of state acquisition, estimation, optimization, and feedback, effectively improving the power response, energy utilization, and operation safety of the narrow lane operation platform in complex terrain.

[0051] Embodiment 2 provides a specific application case, which describes the actual application process of the power distribution method of the hybrid operation platform for narrow lane operation in the 10kV distribution line maintenance operation in a certain hutong in Dongcheng District, Beijing. The width of the hutong is 1.8 meters, the length is about 350 meters, the ground is paved with blue bricks, there are several slight slope changes, the maximum slope is 8 degrees, and part of the road is uneven due to long-term disrepair. The operation task is to replace the insulator of a 120-meter-long 10kV overhead line, the operation height is 8.5 meters, there are two operation personnel and tools and materials in the working bucket, and the total load is about 135 kilograms.

[0052] Embodiment 2 performs the power distribution method according to the following specific steps. First, at the beginning of the operation, the hybrid operation platform enters the hutong from the entrance, the IMU sensor continuously collects the vehicle inclination data at a frequency of 100Hz, and the front and rear inclination angles are 1.2 degrees, the left and right inclination angles are 0.8 degrees, all within the low inclination range. The load sensor collects the working bucket load of 135 kilograms, which is in the medium load range. The four wheel speed sensors collect the speeds of the front left wheel 420rpm, the front right wheel 425rpm, the rear left wheel 418rpm, and the rear right wheel 422rpm, and the average speed is calculated is 421 rpm, which is in the slow range. The fuzzy preprocessing unit divides the pitch angle into the low fuzzy set, the load into the medium fuzzy set, and the speed into the slow fuzzy set according to these data. By querying the 7th rule in the fuzzy rule base, i.e. when the pitch angle is low, the load is medium, and the speed is slow, the activation degree of this rule is obtained as 0.85 by minimum operation, and the corresponding output value is 0.35. Since only one rule is fully activated at this time, the uncertainty factor is directly output by the gravity method 0.35, indicating that the terrain is relatively certain.

[0053] The particle filter estimation unit then starts working, initializing 100 particles, and setting the initial state vector of each particle to wheel slip 0.08, diesel engine speed 800 rpm, electric pump power 0 kW, and battery power 0.82. In the prediction stage, according to the current control input, i.e. the diesel engine throttle opening is 15% and the electric pump is in the off state, the predicted state of each particle after 0.5 seconds is calculated by the nonlinear state transition function. The state transition function is based on the vehicle longitudinal dynamics model, taking into account the tire-ground friction characteristics and the transmission system characteristics. In the weight update stage, the sensor observation values, i.e. the average wheel speed 421 rpm and the battery current 0.5 A, are obtained, and the likelihood probability of each particle with the observation values is calculated. For example, the predicted wheel speed of the 23rd particle is 425 rpm, which is 4 rpm different from the observation value, and the likelihood probability is calculated as 0.92. The weight of this particle is updated from the initial 0.01 to 0.0092 and normalized. After the weight update and normalization of the 100 particles, the effective particle number is 78, which is greater than the threshold value 50, so this iteration does not perform resampling. The wheel slip estimation value is 0.082, and the energy surplus estimation value is 1.37, where the diesel tank fuel percentage is 0.55 and the battery power percentage is 0.82.

[0054] Next, the MPC predictive control unit performs multi-stage prediction allocation. The prediction time domain is set to 10 steps, with a sampling time of 0.5 seconds per step, and the state evolution in the next 5 seconds is predicted. According to the current state and the prediction model, a quadratic objective function is constructed. In the first energy consumption term, the total energy consumption at the current time is 0.18 kWh by integrating the instantaneous power, and the reference energy consumption is set to 0.15 kWh, and the deviation square is 0.0009. In the second pitch angle term, the current pitch angle is 1.2 degrees, and the pitch angle square is 1.44. In the third slip rate term, the current slip rate is 0.082, and the square of the slip rate is 0.0067. In the control term, the power at the last time is 2.8 kW, the current power is 3.2 kW, and the power change rate is 0.4 kW, and the square of the change rate is 0.16. These terms are weighted and summed according to the weight coefficients 0.6, 0.2, 0.2, 0.1 to obtain the target function value at the current time 0.0005 + 0.288 + 0.0013 + 0.016 = 0.3058. This calculation is repeated for the next 10 steps to obtain the total target function value 2.847.

[0055] In terms of constraints, the diesel power plus the electric power equals the total power demand . At the current time, since the vehicle is in a low-speed walking state and the boom is not extended, the power demand is only 3.2 kW. The inequality constraints limit the diesel power to be between 0 and 35.9 kW, the electric power to be between 0 and 15 kW, and the battery power to be between 0.2 and 0.95, and the current state satisfies the constraint conditions. The sequential quadratic programming method is used to solve the optimization problem, and after 8 iterations, the calculation converges in 6.8 milliseconds. The optimal power allocation sequence for the next 10 steps is obtained, and the first step is taken as the power allocation instruction at the current time, i.e., the diesel power allocation is 2.56 kW, the electric power allocation is 0.64 kW, and the hydraulic power allocation is 0 kW. The multi-stage allocation takes 80% of the diesel power, i.e., 2.05 kW, as the basic power, 0.64 kW of electric power as the compensation power, and 0 kW of hydraulic power as the peak buffer.

[0056] In the adaptive feedback and switching phase, the actual output power of the actuator is read every 0.1 seconds. The actual output power of the diesel engine is 2.51 kW, and the actual output power of the electric pump is 0.68 kW, and their sum is 3.19 kW, which has an error of 0.01 kW compared to the power demand of 3.2 kW, and the absolute value of the error is less than the threshold of 0.5 kW, so the PID adjustment is not triggered. The uncertainty factor is 0.35, which is less than the threshold of 0.8, so the system remains in the normal mode and does not switch to the conservative mode. The inclination is 1.2 degrees, and is 0.8 degrees, both of which are less than 5 degrees, so the inclination protection is not triggered. The load No torque protection is triggered for a vehicle mass of 135 kg less than 150 kg or a working range of 0 m less than 5.6 m. No energy recovery is initiated as the vehicle is detected to be not braking.

[0057] When the working platform travels to the middle of the alley about 160 m, it encounters an uphill road section with a slope of 6.5 degrees. The IMU sensor detects the front and rear inclination angles increased to 6.8 degrees, and the left and right inclination angles 1.3 degrees. The load remains 135 kg. The wheel speed sensor detects the average rotation speed rises to 720 rpm to cope with the climbing resistance. The fuzzy preprocessing unit classifies the inclination angle into the high fuzzy set, the load into the medium fuzzy set, and the rotation speed into the medium fuzzy set. The 14th rule in the fuzzy rule base is activated, which indicates that when the inclination angle is high, the load is medium, and the rotation speed is medium, the uncertainty factor should be higher, the rule output value 0.75, the activation degree is 0.88, and the output after defuzzification is 0.66. The particle filter estimation unit updates the state estimation, and the tire slip ratio estimation value is updated to 0.125, indicating that the tire has a certain degree of slip. The energy surplus is updated to 1.28, in which the diesel tank oil percentage drops to 0.52, and the battery power percentage drops to 0.76.

[0058] The MPC predictive control unit re-optimizes the power distribution according to the updated state. Due to the increase in inclination angle and slip ratio, the inclination angle term and slip ratio term weights in the target function increase, and the optimizer tends to output more power to quickly pass through the slope section and reduce slip. The optimization solution is that the diesel power distribution is 14.5 kW, the electric power distribution is 3.8 kW, and the total power demand is 18.3 kW. The adaptive control unit judges that the uncertainty factor 0.66 is less than the threshold value 0.8, and still remains in normal mode. But it judges that the inclination angle 6.8 degrees is greater than 5 degrees, triggering the inclination protection, sending a restriction signal to the control system, prohibiting the extension action of the main telescopic arm and the amplitude action of hydraulic mechanism I. At this time, the working platform only performs walking action, and the arm is kept in the retracted state, so this restriction does not affect the current work.

[0059] When the working platform drives through the slope section to the work point, the vehicle stops on the flat ground, and the inclination angle returns to 1.5 degrees, is 0.9 degree. The operator starts to operate the boom to extend and luff. The main boom extends from 3 meters to 8.5 meters, and the insulated boom is lifted from the horizontal position to 55 degrees with the horizontal plane. During the boom extension, the hydraulic system power demand increases significantly, and the instantaneous power demand reaches 28.6 kW. The IMU sensor detects that the vehicle center of gravity moves forward due to the boom extension, and the front and rear inclination angles increase to 3.8 degrees. The load sensor detects that the working bucket load remains 135 kg. The wheel speed sensor detects that the vehicle is stationary, and the average rotation speed is 0 rpm. The fuzzy preprocessing unit divides the inclination angle into the medium fuzzy set, the load into the medium fuzzy set, the rotation speed into the slow fuzzy set, and the output uncertainty factor is 0.42.

[0060] The particle filter estimation unit estimates that the wheel slip rate converges to 0.05, indicating that the wheel and the ground maintain good adhesion. The energy margin is 1.21, where the diesel tank oil volume percentage is 0.48, and the battery power percentage is 0.73. The MPC predictive control unit faces a large power demand of 28.6 kW, and the optimization solution obtains the diesel power distribution of 22.9 kW accounting for 80% of the total demand, the electric power distribution of 5.7 kW accounting for 20% of the total demand, and the hydraulic power distribution of 0 kW. The diesel engine rotation speed increases from idle speed 800 rpm to 2100 rpm, the output torque is 104 Nm, and the power is 22.9 kW. The electric pump rotation speed increases from 0 rpm to 2280 rpm, and the output power is 5.7 kW. The adaptive control unit reads the actual output power, the diesel engine actually outputs 22.6 kW, the electric pump actually outputs 5.9 kW, and the total output is 28.5 kW, which is 0.1 kW different from the demand 28.6 kW, and the absolute error value is less than the threshold, which does not trigger the PID adjustment.

[0061] During the operation, the operator moves the position in the working bucket to adjust the operation posture, causing the load distribution to change. The load sensor detects that the load increases from 135 kg to 142 kg instantaneously, and the operation amplitude remains at 8.5 meters. The fuzzy preprocessing unit updates the load to the heavy fuzzy set, and the output uncertainty factor increases to 0.58. The MPC predictive control unit adjusts the power distribution according to the load increase, and the optimization solution obtains the diesel power distribution of 23.8 kW, the electric power distribution of 6.1 kW, and the total power increases to 29.9 kW to cope with the larger load. The adaptive control unit judges that the load For 142 kg less than 150 kg and working range of 8.5 meters greater than 5.6 meters, but because the load does not reach 150 kg, it does not meet the two conditions at the same time, so the torque protection is not triggered, allowing the work to continue. When the work is completed, the operator retracts the boom and prepares to leave the work point. The boom is retracted from the stretched state, and the insulating arm is lowered from the raised state. During the boom retraction process, the hydraulic cylinder changes from the extension action to the retraction action, and the hydraulic energy stored in the hydraulic system can be utilized. After detecting the boom retraction signal, the energy recovery controller switches the electric pump to the power generation mode. The high-pressure oil in the hydraulic circuit drives the hydraulic motor of the electric pump to reverse, and the permanent magnet synchronous motor also reverses and generates an induced electromotive force, outputting electric energy to charge the battery pack. During the 5-second retraction process, the battery current sensor detects that the average charging current is 8.2A, and the recovered electric energy is calculated by integration 8.2A multiplied by 5 seconds divided by 3600 seconds per hour, which is 0.0114 Ah, i.e. 11.4 mAh. The battery percentage increases from 0.73 to 0.74, and the energy recovery efficiency is about 35%. The work platform drives away from the work point and returns to the alley entrance along the original route. When passing through the previous uphill section on the way back, it is now in a downhill direction. The operator controls the vehicle to slow down and downshift, and steps on the brake pedal. The brake pedal switch is closed, and a low-level brake signal is output. The energy recovery controller detects the brake signal, closes the pumping function of the electric pump, and switches the electric pump to the power generation mode. The kinetic energy of the vehicle is transmitted to the hub motor through the transmission system, and the hub motor is switched to the generator mode, generating electric energy to charge the battery pack. During the 12-second braking process, the average charging current is 12.5A, and the recovered electric energy is 12.5A multiplied by 12 seconds divided by 3600 seconds per hour, which is 0.0417 Ah, i.e. 41.7 mAh. The battery percentage increases from 0.74 to 0.76, and the energy recovery efficiency is about 42%.

[0062] The whole operation process lasted 65 minutes, the total travel distance was 700 meters, and the total operation time included walking, boom operation and personnel operation. The cumulative working time of the diesel engine was 58 minutes, the average output power was 12.3kW, and the diesel consumption was 3.2 liters. The cumulative working time of the electric pump was 32 minutes, the average output power was 4.8kW, the battery power consumption decreased from 0.82 initially to 0.76 finally, the net consumption was 0.06, which was 6% of the power, considering the battery capacity of 100Ah and the voltage of 48V, the power consumption was 100Ah multiplied by 0.06 multiplied by 48V divided by 1000, which was 0.288kWh, that is, 288Wh. The total energy consumption was the sum of the diesel energy consumption and the electric energy consumption. The diesel energy consumption was 32.96kWh, calculated according to the diesel heat value of 10.3kWh per liter, and the total energy consumption was 32.96 plus 0.288, equal to 33.248kWh. The standard deviation of the vehicle inclination fluctuation throughout the operation was calculated by taking the root mean square deviation of the 78000 inclination data collected in 65 minutes, and the standard deviation of the inclination fluctuation was 1.32 degrees. The maximum value of the wheel slip rate in the whole operation process appeared on the uphill section, which was 0.125, and the average value was 0.068. The power fluctuation was calculated by calculating the root mean square value of the power difference between adjacent sampling time, and the power fluctuation was 2.8kW.

[0063] Comparative Example 1 uses a traditional rule-based power allocation method. This method does not use fuzzy logic, particle filter and model predictive control, but uses a simple threshold judgment and fixed proportion allocation strategy. The power allocation rule is to use only the diesel engine when the power demand is less than 15 kW, to use the diesel engine to provide 70% power and the electric pump to provide 30% power when the power demand is between 15 kW and 30 kW, and to use the diesel engine to provide maximum power 35.9 kW and the electric pump to make up the rest when the power demand is greater than 30 kW. The hidden state estimation uses a simple moving average filter, the wheel slip ratio is calculated by the difference between the wheel speed and the vehicle speed divided by the vehicle speed, and the tire dynamics and sensor noise are not considered. The terrain adaptability is not quantified, and only the vehicle speed is reduced when the inclination exceeds 5 degrees, but the power allocation ratio is not adjusted. Comparative Example 1 performs the same work task in the same work scenario. In the alley entrance stage, the power demand is 3.2 kW, which is less than 15 kW, the rule judges to use only the diesel engine, the electric pump is turned off, the diesel power is 3.2 kW, and the electric power is 0 kW. In the uphill section stage, the power demand is 18.3 kW, which is in the 15 kW to 30 kW interval, the diesel power is 12.81 kW, which is 70%, and the electric power is 5.49 kW, which is 30%. In the boom stretching stage, the power demand is 28.6 kW, the diesel power is 20.02 kW, which is 70%, and the electric power is 8.58 kW, which is 30%. Due to the lack of prediction ability, the response lags behind when the power demand suddenly changes, resulting in instantaneous power shortage or excess. In the uphill section, due to the non-optimized power allocation, the low output power of the diesel engine leads to a decrease in vehicle speed and more serious tire slip, with a maximum wheel slip ratio of 0.187. The simple moving average filter has insufficient estimation accuracy in a nonlinear environment, and the wheel slip ratio estimation value fluctuates greatly, with an average estimation error of 0.035. Since there is no energy recovery mechanism, the energy in the boom retraction and braking process is completely wasted and does not charge the battery. The entire work process of Comparative Example 1 lasts 69 minutes, which is 4 minutes longer than Example 2, mainly because the slow vehicle speed and power response lag in the uphill section lead to a decrease in work efficiency. The cumulative working time of the diesel engine is 64 minutes, the average output power is 13.8 kW, and the diesel consumption is 4.1 liters. The cumulative working time of the electric pump is 28 minutes, the average output power is 6.2 kW, the battery power consumption decreases from 0.82 to 0.68, the net consumption is 0.14, which is 14% of the power, and the power consumption is 0.672 kWh, which is 672 Wh. The total energy consumption is 4.1 liters of diesel multiplied by 10.3 kWh per liter plus 0.672 kWh, which is equal to 42.23 plus 0.672, which is equal to 42.902 kWh. The standard deviation of the inclination fluctuation is 2.15 degrees, which is 63% larger than Example 2, indicating that the vehicle stability is poor. The maximum wheel slip ratio is 0.187, and the average value is 0.102, both of which are higher than Example 2. The power fluctuation is 4.7 kW, which is 68% larger than Example 2, indicating that the power change is more intense.

[0064] Comparative Example 2, which uses extended Kalman filter instead of particle filter for hidden state estimation, is the same as Example 2 except for the hidden state estimation. The extended Kalman filter is a first-order Taylor expansion linearization of the nonlinear system around the operating point, and then the standard Kalman filter algorithm is applied. This method assumes that the system noise and observation noise are subject to Gaussian distribution, and the estimation accuracy is limited under the strong nonlinearity and non-Gaussian characteristics of the narrow lane environment. The state transition function calculates the partial derivative through the Jacobian matrix, and the observation function is also linearized through the Jacobian matrix. The Kalman gain is calculated by the state covariance matrix and the observation noise covariance matrix. In the same operating scenario, the wheel slip ratio estimation value at the entrance of the alley stage of Comparative Example 2 is 0.085, and the error compared with 0.082 of Example 2 is 0.003, and the error is smaller because the system is in a stable road section at this time. The degree of nonlinearity is low. In the uphill road section stage, due to the strong nonlinearity of the tire adhesion characteristics, the linearization assumption of the extended Kalman filter is no longer established, and the wheel slip ratio estimation value is 0.148, and the error compared with the actual value 0.125 is 0.023, while the particle filter estimation value of Example 2 is 0.125, and the error is 0. In the boom extension stage, the wheel slip ratio estimation converges to 0.06 when the vehicle is stationary, and the error compared with the actual value 0.05 is 0.01, and the error of Example 2 is 0. Due to the inaccurate hidden state estimation, the MPC predictive control unit optimizes based on the wrong state information, resulting in a power distribution that is not a truly optimal solution. In the uphill section, due to the overestimation of the wheel slip ratio, the optimizer allocates too much power to try to reduce the slip, and the actual power allocation is diesel power 16.2kW, electric power 4.3kW, and total power 20.5kW, which exceeds the actual demand of 18.3kW, resulting in energy waste. The entire operating process of Comparative Example 2 lasts 66 minutes, which is 1 minute more than Example 2. The cumulative working time of the diesel engine is 59 minutes, the average output power is 13.1kW, and the diesel consumption is 3.5 liters. The cumulative working time of the electric pump is 33 minutes, the average output power is 5.2kW, the battery power consumption decreases from 0.82 to 0.74, the net consumption is 0.08, which is 8% of the power, and the power consumption is 0.384kWh, which is 384Wh. The total energy consumption is 3.5 liters of diesel multiplied by 10.3kWh per liter plus 0.384kWh, which equals 36.05 plus 0.384, which equals 36.434kWh. The standard deviation of the inclination angle fluctuation is 1.58 degrees, which is 20% larger than Example 2. The maximum value of the wheel slip ratio is 0.148, and the average value is 0.081, which are higher than Example 2. The power fluctuation is 3.4kW, which is 21% larger than Example 2.

[0065] Comparative Example 3, which did not use fuzzy logic to process terrain uncertainty, directly input sensor raw data into the MPC predictive control unit, and was otherwise identical to Example 2. Due to the lack of fuzzy processing, the MPC optimizer was unable to quantify the complexity of the terrain, and still followed the strategy for a deterministic environment to allocate power in an uncertain environment. On the uphill section, although the inclination reached 6.8 degrees, indicating a complex terrain, the system did not output an uncertainty factor , and the MPC optimizer did not switch to a conservative mode, still maintaining the electric power ratio at 20%. Due to the fast response speed of the electric pump, increasing the electric power ratio in an uncertain environment helps to cope with sudden load changes. Comparative Example 3 lacks this adaptive ability. On the uphill section, the power allocation of Comparative Example 3 was 14.6 kW of diesel power and 3.7 kW of electric power, slightly higher diesel power and slightly lower electric power compared to Example 2. When encountering sudden road bumps that cause an increase in instantaneous load, the diesel engine cannot increase output in time due to the lag in speed response, while the electric pump lacks power reserves, resulting in an instantaneous power gap of 1.8 kW for 0.3 seconds, causing the vehicle speed to decrease and the tire slip to increase. During the boom extension stage, due to the lack of mode switching mechanism, the system failed to adaptively adjust the power source combination according to the load change, and the power allocation flexibility decreased. The entire operation process of Comparative Example 3 lasted 67 minutes, 2 minutes longer than Example 2. The diesel engine worked for a total of 60 minutes, with an average output power of 12.9 kW, consuming 3.6 liters of diesel. The electric pump worked for a total of 31 minutes, with an average output power of 4.9 kW, consuming 0.07 or 7% of the battery power from 0.82 to 0.75, and consuming 0.336 kWh or 336 Wh of electric energy. The total energy consumption is 3.6 liters of diesel multiplied by 10.3 kWh per liter plus 0.336 kWh, which equals 37.08 plus 0.336, which equals 37.416 kWh. The standard deviation of the inclination fluctuation was 1.65 degrees, 25% larger than Example 2. The maximum wheel slip rate was 0.142, and the average was 0.085, both higher than Example 2. The power fluctuation was 3.6 kW, 29% larger than Example 2.

[0066] Comparative experiment design and result analysis; According to the relevant provisions of national standards GB / T19750-2005 Hybrid electric vehicle energy consumption test methods and GB / T18385-2005 Electric vehicle power performance test methods, install test instruments on the hybrid operation platform to carry out real vehicle road test. The test site is selected in a typical hutong in Dongcheng district of Beijing, the hutong length is 350 meters, the width is 1.8 meters, the road surface is laid with blue bricks, there are slope changes and uneven sections, which meets the actual working conditions of narrow lane operation. The test vehicle is the hybrid operation platform described in embodiment 1, the external dimensions are 6.1 meters long, 1.2 meters wide and 2.7 meters high, the hybrid system includes a 35.9kW diesel engine and a 15kW electric pump. The test instrument configuration is as follows. The energy consumption measurement uses an oil consumption meter and an electric meter, the oil consumption meter model is FC-2210, the precision is ±0.5%, the instantaneous oil consumption and cumulative oil consumption are measured by being connected in series in the fuel pipeline, the sampling frequency is 10Hz. The electric meter model is DW-B103, the precision is ±0.2%, the instantaneous current and cumulative electric quantity are measured by being connected in series at the output end of the battery pack, the sampling frequency is 20Hz. The inclination measurement uses the MPU6050 type IMU sensor in embodiment 1, the precision is ±0.1 degree, the sampling frequency is 100Hz, and the inclination data is transmitted to the data acquisition system through the CAN bus. The wheel slip rate measurement uses a combination of an optical vehicle speed sensor and a wheel speed sensor, the vehicle speed sensor model is SG-100, which is installed at the non-driving position of the chassis to obtain the true vehicle speed by measuring the relative motion speed of the ground, the precision is ±0.5%, and the wheel speed sensor is the Hall type sensor in embodiment 1, the wheel slip rate is obtained by calculating the deviation between the vehicle speed and the wheel speed. The power measurement uses a power analyzer, model PA310, precision ±0.1%, measures the output power of the diesel engine and the output power of the electric pump respectively, sampling frequency 50Hz. All test data is collected to the data acquisition system through the CAN bus, the acquisition system uses Advantech industrial computer, model ARK-1123, equipped with CAN communication interface card and 1TB solid state disk for storing test data. The test working condition is designed to simulate the typical distribution line maintenance operation process, including seven stages of no-load driving, load driving, uphill, boom stretching, overhead operation, boom retracting and downhill braking. The specific parameters of each stage are as follows: no-load driving stage, vehicle speed 2.5 kilometers per hour, duration 8 minutes; load driving stage, load 135 kilograms, vehicle speed 2.0 kilometers per hour, duration 12 minutes; uphill stage, slope 6.5 degrees, load 135 kilograms, vehicle speed 1.5 kilometers per hour, duration 3 minutes; boom stretching stage, stretching from 3 meters to 8.5 meters, duration 4 minutes; overhead operation stage, boom fixed at 8.5 meters, load 135 kilograms, duration 30 minutes; boom retracting stage, retraction from 8.5 meters to 3 meters, duration 3 minutes; downhill braking stage, slope 6.5 degrees, initial speed 2.0 kilometers per hour, duration 2 minutes. The total duration of the whole test working condition is 62 minutes, each group of test is repeated three times to take the average value to eliminate accidental errors.

[0067] The fuel system and battery system are calibrated before the test according to the provisions of GB / T 19750-2005 standard. The fuel tank is filled to the specified level, and the fueling amount is recorded as 60 liters. The battery pack is fully charged to 95% of the electricity percentage, and the system is allowed to reach thermal equilibrium for 30 minutes. During the test, the ambient temperature is controlled at 20±3 degrees Celsius, and the relative humidity is controlled at 50±10%, which meets the test conditions specified in the standard. After the test, the fuel tank level and battery electricity are measured again, and the energy consumption is calculated by the difference before and after. According to the provisions of article 5.3.2 of GB / T 19750-2005 standard, the comprehensive energy consumption of the hybrid electric vehicle is weighted calculated by the heat value of fuel consumption and the electricity consumption. The heat value of diesel is taken as the standard value of 10.3 kilowatt hours per liter, and the electricity is directly calculated according to the measured value.

[0068] In the three repeated tests, the fuel consumption of example 2 is 3.18 liters, 3.22 liters and 3.21 liters, respectively, with an average of 3.20 liters, a standard deviation of 0.021 liters, and a coefficient of variation of 0.66%, indicating good repeatability of the data. The electricity consumption is 0.285 kilowatt hours, 0.291 kilowatt hours and 0.288 kilowatt hours, respectively, with an average of 0.288 kilowatt hours and a standard deviation of 0.003 kilowatt hours. The comprehensive energy consumption is 3.20 liters multiplied by 10.3 kilowatt hours per liter plus 0.288 kilowatt hours, which equals 32.96 plus 0.288, which equals 33.248 kilowatt hours. The average fuel consumption of comparative example 1 is 4.08 liters, the average electricity consumption is 0.668 kilowatt hours, and the comprehensive energy consumption is 42.024 plus 0.668, which equals 42.692 kilowatt hours. The average fuel consumption of comparative example 2 is 3.48 liters, the average electricity consumption is 0.381 kilowatt hours, and the comprehensive energy consumption is 35.844 plus 0.381, which equals 36.225 kilowatt hours. The average fuel consumption of comparative example 3 is 3.58 liters, the average electricity consumption is 0.334 kilowatt hours, and the comprehensive energy consumption is 36.874 plus 0.334, which equals 37.208 kilowatt hours. The comprehensive energy consumption of example 2 is reduced by 22.1% compared with comparative example 1, by 8.2% compared with comparative example 2, and by 10.6% compared with comparative example 3, which proves that the method of the present application has significant advantages in energy consumption.

[0069] The work efficiency is represented by the time required to complete the standard work condition, and the shorter the time, the higher the efficiency. According to the provisions of GB / T18385-2005 standard article 4.2 about dynamic performance test, the total time of the vehicle completing all test conditions is recorded. The average work time of Example 2 is 64.8 minutes, and the standard deviation is 1.2 minutes. The average work time of Comparative Example 1 is 68.5 minutes, which is 3.7 minutes more than Example 2, and the efficiency is reduced by 5.7%, mainly because of the power response lag and slow speed on the uphill section. The average work time of Comparative Example 2 is 65.9 minutes, which is 1.1 minutes more than Example 2, and the efficiency is reduced by 1.7%. The average work time of Comparative Example 3 is 66.8 minutes, which is 2.0 minutes more than Example 2, and the efficiency is reduced by 3.1%. Example 2 performs best in work efficiency, indicating that the optimized power distribution strategy can improve the work response speed.

[0070] The stability index test results are as follows. The stability is represented by the standard deviation of the vehicle inclination angle fluctuation, and the smaller the standard deviation, the better the stability. According to the ISO14982:1998 standard for stability test of agricultural and forestry machinery, the front and rear inclination angles and the left and right inclination angles of the vehicle are continuously recorded during the test, the sampling frequency is 100 Hz, the test time is 62 minutes, and a total of 372000 data points are collected. The root mean square deviation is calculated for each group of data as the inclination angle fluctuation standard deviation. The average value of the front and rear inclination angle fluctuation standard deviation of Example 2 is 1.31 degrees, the average value of the left and right inclination angle fluctuation standard deviation is 0.88 degrees, and the comprehensive inclination angle fluctuation standard deviation is 1.59 degrees. The comprehensive inclination angle fluctuation standard deviation of Comparative Example 1 is 2.14 degrees, which is 34.6% larger than that of Example 2. The comprehensive inclination angle fluctuation standard deviation of Comparative Example 2 is 1.57 degrees, which is slightly smaller than that of Example 2 by 1.3%, but considering the measurement error, they are basically the same. The comprehensive inclination angle fluctuation standard deviation of Comparative Example 3 is 1.64 degrees, which is 3.1% larger than that of Example 2. Example 2 and Comparative Example 2 perform similarly in stability, both of which are better than Comparative Example 1 and Comparative Example 3, indicating that MPC predictive control plays a key role in improving stability.

[0071] Wheel slip ratio reflects the adhesion between the tire and the ground; a lower slip ratio indicates better adhesion. According to Clause 4.4 of GB / T18385-2005 standard regarding the climbing ability test, wheel slip ratio was measured on an uphill section. In Example 2, the maximum average wheel slip ratio on the uphill section was 0.124, and the average value was 0.067. In Comparative Example 1, the maximum average wheel slip ratio was 0.185, 49.2% higher than Example 2, and the average value was 0.101, 50.7% higher than Example 2. In Comparative Example 2, the maximum average wheel slip ratio was 0.146, 17.7% higher than Example 2, and the average value was 0.080, 19.4% higher than Example 2. In Comparative Example 3, the maximum average wheel slip ratio was 0.141, 13.7% higher than Example 2, and the average value was 0.084, 25.4% higher than Example 2. The wheel slip ratio index of Example 2 is significantly better than that of the other pairs, indicating that particle filtering can accurately estimate the wheel slip ratio state and MPC can effectively reduce tire slippage.

[0072] Power fluctuation is represented by the root mean square (RMS) value of the power difference between adjacent sampling times; the smaller the fluctuation, the more stable the power output. Large power fluctuations lead to frequent speed adjustments in the powertrain, increasing mechanical wear and reducing ride comfort. Power data was sampled at a frequency of 50Hz throughout the test, and the RMS value of the power difference between adjacent sampling points was calculated. The average power fluctuation of Example 2 was 2.79 kW. The average power fluctuation of Comparative Example 1 was 4.68 kW, 67.7% larger than Example 2. The average power fluctuation of Comparative Example 2 was 3.38 kW, 21.1% larger than Example 2. The average power fluctuation of Comparative Example 3 was 3.57 kW, 28.0% larger than Example 2. Example 2 exhibited the smallest power fluctuation, indicating that the power change rate penalty term in the MPC optimization objective function effectively smoothed the power output. Test results are shown in Table 1 and... Figures 5-7 .

[0073] Table 1. Comparison of performance indicators between Example 2 and each comparative example.

[0074]

[0075] From Table 1, it can be seen that Example 2 is superior to Comparative Examples 1, 2 and 3 in comprehensive energy consumption, operation efficiency, wheel slip rate and power fluctuation, and is basically equivalent to Comparative Example 2 in stability. The comprehensive energy consumption is reduced by 8.2% to 22.1%, with an average reduction of 13.6%, achieving a significant energy-saving effect. The operation efficiency is increased by 1.7% to 5.7%, with an average increase of 3.5%, which can shorten the operation time and improve the work efficiency. The wheel slip rate is reduced by 13.7% to 50.7%, with an average reduction of 27.5%, which significantly improves the adhesion performance and passing performance of the vehicle on complex terrain. The power fluctuation is reduced by 21.1% to 67.7%, with an average reduction of 38.9%, which greatly improves the stability of power output. These experimental data fully prove the superiority of the method of the application.

[0076] Figure 5 The curves of diesel power, electric power and total power changing with time are shown, and the power data are obtained by measuring the output power of the diesel engine and the electric pump in real time by the power analyzer PA310 at a sampling frequency of 50 Hz. Figure 5 Reflects the stability and response characteristics of the power distribution strategy. From the subgraph of Figure 5 It can be seen from the subgraph of that the power curve of Example 2 is smooth and continuous, and the transition is natural when the working condition is converted. The diesel power and the electric power are coordinated, especially in the uphill stage, the diesel power is smoothly increased from 8.5 kW to 14.5 kW, and the electric power is continuously increased to 22.9 kW in the boom extension stage, and there is no obvious mutation in the whole process. The power curve of Comparative Example 1 presents a step-shaped jump, especially in the uphill stage at 20.5 minutes, the diesel power suddenly jumps from 10.6 kW to 12.8 kW, and the electric power suddenly jumps from 0 kW to 5.5 kW, which will cause mechanical impact and energy waste. The power of Comparative Example 2 reaches 20.5 kW in the uphill stage, which is about 2.2 kW more than the actual demand of 18.3 kW, which shows that the power distribution is too conservative due to the inaccurate estimation of the extended Kalman filter in the nonlinear environment. Comparative Example 3 has a power gap when the instantaneous load changes, which shows that the lack of terrain self-adaptive ability of fuzzy logic leads to insufficient power reserve. It shows that the MPC predictive control of Example 2 realizes smooth transition by rolling optimization of the future 10-step power distribution sequence, avoids power mutation, and the wheel slip rate and energy margin accurately estimated by the particle filter provide reliable state information support for MPC to optimize the decision.

[0077] Figure 6The wheel slip rate curve of the uphill section 180 seconds is shown, the wheel slip rate is calculated by the difference between the true vehicle speed measured by the photoelectric speed sensor SG-100 and the wheel speed measured by the wheel speed sensor divided by the vehicle speed, which reflects the tire and ground adhesion and the degree of slip. The wheel slip rate curve of Example 2 fluctuates the least, the peak value is 0.124 appearing at about 90 seconds in the middle of the uphill, the average value is 0.067, and the whole curve presents a smooth single peak shape. The wheel slip rate curve of Comparative Example 1 fluctuates violently, the peak value reaches 0.185, which is about 49% higher than Example 2, the average value is 0.101, which is about 51% higher than Example 2, and the curve presents irregular sawtooth fluctuation. The peak value of the wheel slip rate of Comparative Example 2 is 0.146, which is about 18% higher than Example 2, the average value is 0.080, which is about 19% higher than Example 2, and the fluctuation degree is between Example 2 and Comparative Example 1. The peak value of the wheel slip rate of Comparative Example 3 is 0.141 and the average value is 0.084, which is slightly better than Comparative Example 2 but still significantly worse than Example 2. It shows that the particle filter searches in the nonlinear state space through 100 particles, accurately estimates the true value of the wheel slip rate, provides accurate state feedback for MPC, so that MPC can optimize power distribution according to the wheel slip rate, and timely increase power output to reduce slip when detecting that the wheel slip rate increases. The simple moving average filter of Comparative Example 1 and the extended Kalman filter of Comparative Example 2 have larger estimation errors in the strong nonlinear environment, which leads to improper power distribution and aggravates tire slip.

[0078] Figure 7 The comprehensive energy consumption of four methods to complete the same task is compared intuitively through the histogram, which is calculated according to the difference between the fuel tank level before and after the test multiplied by the diesel heat value 10.3 kWh per liter, plus the difference in battery capacity. The comprehensive energy consumption of Example 2 is 33.248 kWh, which is the lowest among the four methods, Comparative Example 1 is 42.692 kWh, which is 28.4% higher than Example 2, Comparative Example 2 is 36.225 kWh, which is 9.0% higher than Example 2, and Comparative Example 3 is 37.208 kWh, which is 11.9% higher than Example 2. The energy saving effect of Example 2 comes from the MPC predictive control which optimizes the energy consumption term in the objective function to achieve global energy optimal distribution, avoiding local suboptimality of traditional methods, while the fuzzy logic adjusts the ratio of electric power and diesel power according to the terrain uncertainty factor, increases the electric power ratio to 60% in the conservative mode to reduce the diesel engine speed and reduce fuel consumption. In addition, the energy recovery mechanism recovers 11.4 mAh of electric quantity in the boom retraction stage and 41.7 mAh of electric quantity in the braking stage, with a total recovery of 53.1 mAh, which is equivalent to 0.053% of the battery capacity. Although the absolute value is not large, it reflects the secondary utilization of energy.

[0079] The method of the present application provides terrain self-adaptive ability for power distribution by converting complex terrain features difficult to quantify into quantitative uncertainty factors through fuzzy logic, the particle filter overcomes the limitations of traditional filtering methods in nonlinear non-Gaussian environment and accurately estimates the wheel slip rate and energy balance to provide a reliable basis for optimization decision, and the MPC realizes global optimal energy management by predicting future state evolution and rolling optimization control sequence, and the three work together to realize the comprehensive advantages of energy consumption reduction, efficiency improvement, stability enhancement, adhesion performance improvement and power output stability.

[0080] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A hybrid work platform power distribution method for narrow aisle operations, characterized by, Comprising the following steps: S1: real-time state acquisition and fuzzy preprocessing: collect vehicle inclination, acceleration, working bucket load and four-wheel rotation speed data; divide the inclination data into three fuzzy sets of low, medium and high, divide the load data into three fuzzy sets of light, medium and heavy, and divide the rotation speed data into three fuzzy sets of slow, medium and fast; process three input fuzzy sets through a fuzzy logic controller to output an uncertainty factor ; S2: Particle filter hidden state estimation: initialize particle swarm; based on the collected sensor data, estimate the wheel slip rate through the particle prediction, weight update and resampling process and energy reserves ; S3: MPC multi-stage prediction distribution: set the prediction time domain and sampling time; construct a target function containing energy consumption items, inclination items and slip rate items; set power balance constraints and battery power constraints; solve the quadratic programming optimization problem, output the first-stage diesel power distribution, the second-stage electric power distribution and the third-stage hydraulic peak buffer distribution; S4: Adaptive feedback and switching: Perform feedback control; Determine uncertainty factor Whether greater than threshold, if greater than threshold, switch to conservative mode to increase electric power ratio; Determine whether the inclination and load are out of limits, if out of limits, trigger protection limit arm action; Start energy recovery to charge the battery when braking.

2. The method of claim 1, wherein, The S1 specifically comprises: S11: Collect the front and rear inclination angles of the vehicle through the IMU sensor at a frequency of 100 Hz and the left and right inclination angles in degrees; when or , set the inclination angle alarm flag bit to 1; S12: Collect the current load weight of the working bucket through the load sensor at a frequency of 50Hz , in kilograms; Collect the front left wheel speed , front right wheel speed , rear left wheel speed , and rear right wheel speed of the four wheel speed sensors, in revolutions per minute; S13: defining a membership function according to the inclination angle , , when belongs to the "low" fuzzy set, when belongs to the "medium" fuzzy set, and when belongs to the "high" fuzzy set; defining a membership function according to the load , when belongs to the "light" fuzzy set, when belongs to the "medium" fuzzy set, and when belongs to the "heavy" fuzzy set; defining a membership function according to the average rotational speed , when belongs to the "slow" fuzzy set, when belongs to the "medium" fuzzy set, and when belongs to the "fast" fuzzy set; wherein is the inclination angle value; is the load value; is the average rotational speed; , , , are the four-wheel rotational speeds, respectively. S14: Establish fuzzy rule base, including 27 rules; input membership values of three fuzzy sets, solve the fuzzy by the gravity method, calculate the output uncertainty factor ; wherein is the activation degree of the th rule; is the output value corresponding to the th rule; represents the summation operator; convert the fuzzy terrain information into quantitative uncertainty factor.

3. The method of claim 1, wherein, The S2 specifically comprises: S21: Initialization There are 10 particles, and the initial state vector of each particle is 10. ,in The initial wheel slip ratio, This is the initial diesel engine speed. This is the initial electric pump power. Set the initial battery level; assign initial weights to each particle. Set as ;in The total number of particles; For particle index; superscript Indicates transpose; S22: Perform state prediction for each particle, the prediction formula is ; wherein is the predicted state of the i-th particle at the j-th time instant; is a nonlinear state transition function; is the control input vector at the j-th time instant, including the diesel engine throttle opening and the electric pump start-stop signal; is the process noise, which is subject to a Gaussian distribution with a mean of 0 and a variance of ;​​​ S23: Obtain sensor observations ; Calculate likelihood probability of each particle ; Update weight of each particle ; Normalize weights ; Where is the observation vector at time ; is the observed average wheel speed; is the observed battery current; is the observation function; is the observation noise covariance; denotes the Euclidean norm of a vector; is the natural exponential function; is the summation index; S24: Calculate the number of effective particles When , perform resampling: resample particles according to the size of weights , retain high-weight particles, and eliminate low-weight particles; after resampling, reset all weights to ; calculate the wheel slip rate estimate and the energy surplus estimate ; wherein is the number of effective particles; is the diesel tank fuel percentage; is the battery pack power percentage; accurately estimate the hidden state variables in a narrow lane nonlinear environment.

4. The method of claim 1, wherein, The S3 specifically comprises: S31: set prediction horizon sampling time seconds; according to the diesel engine torque characteristic curve, the electric pump response characteristic and the pressure-flow relationship of hydraulic mechanism I and hydraulic mechanism II, a discrete-time prediction model is established ; wherein is the prediction step number; is the sampling interval; is the state transition matrix; is the control input matrix; is the state vector of the first step; is the control vector of the first step; S32: Constructing a quadratic objective function ; wherein is the objective function value; is the total energy consumption of the first step, in units of kilowatt-hours; is the reference energy consumption; is the vehicle inclination angle of the first step; is the wheel slip rate of the first step; is the power change rate; , , is the state weight coefficient; is the control weight coefficient; S33: Establish equality constraints and inequality constraints , , ; wherein is the power of the diesel engine; is the power of the electric pump; is the total power demand; is the charge of the battery bank; kW is kilowatt unit; S34: solve the optimization problem by using the sequence quadratic programming method to obtain the power allocation sequence of the future step ; take the optimization result of the first step , as the power allocation instruction at the current time; wherein the superscript represents the optimal value; the first-level allocation takes 80% of as the basic power, the second-level allocation takes as the compensation power, and the third-level allocation takes as the peak buffer power; the energy consumption is optimized and the stability is best under the condition of meeting the job requirements.

5. The method of claim 1, wherein, The S4 specifically comprises: S41: reading the actual output power of the diesel engine every second and the actual output power of the electric pump ; calculating a power tracking error ; when , adjusting the throttle opening of the diesel engine or the rotation speed of the electric pump through a PID controller in the control system; wherein is a feedback period; is a power error; denotes an absolute value operator; S42: determining the uncertainty factor whether greater than a threshold value ; if , switch to a conservative mode: set the target power ratio of the electric pump to 60%, and the target power ratio of the diesel engine to 40%; the diesel engine runs at a reduced speed in the conservative mode, reducing noise and emissions in a narrow lane environment; wherein is an uncertainty threshold value; S43: judging the vehicle inclination angle or whether ; if yes, sending a restriction signal to the control system, prohibiting the boom extending action of the main telescopic arm and the luffing action of the hydraulic mechanism I by the control box, and only allowing the arm retracting and lowering actions; judging the working bucket load whether and whether the working range is meters; if both conditions are met, triggering the torque protection, and prohibiting all the extending actions of the hydraulic mechanism I and the hydraulic mechanism II by the control system; S44: detecting a vehicle braking signal; when braking is detected, the pumping function of the electric pump is turned off by the control system, the electric pump is switched to the power generation mode; the hydraulic energy in the hydraulic circuit is used to drive the electric pump to generate electricity, the generated electricity is used to charge the battery pack; record the recovered electricity ; wherein is the recovered electricity, in ampere-hours; is the charging current; represents the integral operator; is the time differential; ensure the safety of narrow lane operation and improve energy utilization efficiency.

Images