Anti-disturbance method for electric power material transportation system based on feedforward compensation and robust control

Abstract

The present application relates to the technical field of power energy material supply chain and transportation system control, in particular to an anti-disturbance method for power material transportation system based on feedforward compensation and robust control, comprising: defining controllable state variables and reference steady state at the planning level to construct a unified error signal interface; embedding multi-dimensional mathematical models to realize quantitative control by feedforward compensation of disturbance observer (DOB), smooth adjustment of improved PID control, and rigid constraint of sliding mode variable structure control (SMC); mapping the control output to the planning decision variable through a multi-time scale closed-loop update mechanism, dynamically balancing the system planning stability, anti-disturbance ability and response efficiency without intervening the physical execution layer. The present application is applicable to complex scenarios such as multi-modal intermodal transport, cross-regional allocation and emergency repair of power materials, and can effectively deal with multi-source disturbances and emergencies such as extreme weather and limited loading and unloading in reservoir area, significantly improving the stability and reliability of power grid material supply chain operation.

Description

Technical Field

[0001] This invention relates to the field of power energy material supply chain and transportation system control technology, and in particular to an anti-disturbance control method applicable to multimodal transport and complex dispatching networks of power materials (such as large main grid equipment, distribution network emergency materials, etc.), specifically an anti-disturbance method for power material transportation systems based on feedforward compensation and robust control. Background Technology

[0002] As power grids develop towards intelligence, global coverage, and high reliability, multimodal transport, cross-regional dispatch, and emergency repair transportation of power materials (transformer equipment, cables, fittings, repair equipment, emergency power supplies, etc.) are becoming increasingly common. The disturbance factors faced by the system in operation are becoming more and more complex, including extreme weather (typhoons, blizzards, ice storms), sudden power grid repair tasks, traffic control, insufficient storage capacity, delays in loading and unloading nodes, temporary interruptions in transportation capacity, and emergency requisition of goods for line repairs, among other uncertainties.

[0003] However, existing methods for mitigating disturbances in power material transportation systems have the following main drawbacks:

[0004] On the one hand, the controlled objects focus on parameters such as position and speed at the physical layer, without taking into account the decision-making needs at the material allocation planning level, or the decision-making needs at the power material planning level such as scheduling, emergency repair priority, and supply sequence. On the other hand, there is a lack of quantitative mathematical model support, resulting in poor connection between control signals and scheduling decisions. This can easily lead to frequent rescheduling, failure of emergency repair paths, and frequent replanning. Multiple disturbances directly affect the decision-making level, and there is a lack of a unified error feedback interface and mathematical description. Conflicts can easily arise between different control strategies, reducing control efficiency.

[0005] On the other hand, the mathematical models of single control strategies (such as traditional PID and simple sliding mode control) have poor adaptability and are difficult to cover disturbances of different intensities: PID control has weak anti-disturbance ability and cannot converge quickly when faced with strong disturbances; although sliding mode control is robust, it has chattering problems and lacks quantitative basis for parameter design; the control timing lacks mathematical coordination rules, and it is difficult to balance response speed and stability. Either the response is lagging and cannot suppress disturbances in time, or the response is excessive and causes system oscillation.

[0006] Therefore, there is an urgent need for an anti-disturbance control method that is based on a sound mathematical model, takes into account the needs of power material allocation decision-making, adapts to disturbances of varying intensities, and achieves time-series coordination, so as to improve the operational stability and reliability of the power material transportation system in complex and uncertain environments. Summary of the Invention

[0007] The purpose of this invention is to provide a disturbance mitigation method for power material transportation systems based on feedforward compensation and robust control. By constructing a hierarchical control architecture of "feedforward compensation-smooth adjustment-rigid constraint", embedding a multi-dimensional mathematical model to achieve quantitative control, and combining it with a multi-timescale closed-loop collaborative mechanism, this invention addresses the shortcomings of existing disturbance mitigation methods for power material transportation, such as poor coordination between scheduling and control, scheduling conflicts caused by multi-source disturbances, poor adaptability of single control strategies, and imbalance between response and supply stability. Furthermore, these methods lack system mathematical model support and cannot effectively cope with complex multi-source disturbances and emergencies such as power grid repair and extreme disasters.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0009] A disturbance rejection method for a power material transportation system based on feedforward compensation and robust control includes the following steps:

[0010] S1. Define the controllable state variables and reference steady state of the power material transportation system: obtain the state variables used for feedback control in the power material transportation system, construct the system state vector describing the deviation at the planning level, and determine the reference steady state of the system state vector through the sliding window statistical method or the initial solution of the planning.

[0011] The deviation includes path stability deviation, node disturbance cumulative deviation, and the degree to which the overall system operation deviates from steady state;

[0012] S2. Constructing the system error signal: Based on the difference between the current system state vector and the reference steady state, the three deviations—path stability deviation, cumulative node disturbance deviation, and the degree of overall system operation deviating from the steady state—are dynamically weighted to generate a unified system error signal. The error signal does not distinguish the source of the disturbance, providing a unified feedback interface for different control modules;

[0013] S3. Real-time Estimation and Feedforward Compensation of Unknown Disturbances: A third-order extended state observer is embedded in the system feedback channel as a disturbance observer. By comparing the difference between the nominal system model and the actual state response, the equivalent disturbance, including external environmental disturbances and internal model uncertainties, is estimated in real time. The combined effects of the forms are considered, and the equivalent disturbance is applied to the control input of the power material transportation system through feedforward compensation, so as to suppress the amplified impact of the disturbance on the timeliness of power material transportation and emergency repair guarantee in advance.

[0014] S4. Improved PID control for adjusting system deviation: Based on the system error signal and equivalent disturbance, the compensated error signal is confirmed, and a PID controller with parameter self-tuning is used to continuously and smoothly adjust the deviation of the power material transportation system; the proportional term suppresses the trend of deviation expansion, the integral term eliminates long-term accumulated deviation, and the derivative term suppresses system oscillation; the PID controller acts on the planning layer parameter update frequency and adjustment range;

[0015] S5. Constructing the sliding surface and defining the stability boundary: based on systematic errors. and its rate of change Constructing the sliding surface The sliding surface does not depend on an accurate system model and describes the dynamic trajectory of the power material transportation system converging from an unstable state to a stable state by setting switching rules.

[0016] S6. Sliding mode variable structure control to cope with strong disturbances: when the sliding surface state quantity or disturbance intensity When the effective range is exceeded, sliding mode variable structure control (SMC) is activated. The sliding mode variable structure control enhances the robust control effect by adaptively switching the control law, so that the system state converges quickly along the sliding surface.

[0017] Among them, due to equivalent perturbation Since it can be positive or negative, the absolute value of the equivalent disturbance is taken to obtain the disturbance intensity.

[0018] S7. Mapping control output to planning decision variables: Based on the feedforward compensated control input, the parameterized self-tuning PID controller, and the sliding mode variable structure control analysis, the control output is synthesized. This is then transformed into executable decision adjustment quantities for the power material transportation planning layer through a mapping function. These decision adjustment quantities include path weight corrections. Node avoidance factor adjustment and replanning trigger threshold update ;

[0019] S8. Closed-loop update and multi-timescale collaborative operation: Steps S1-S7 are continuously executed in a closed loop, and state updates are achieved through iterative formulas. The disturbance observer (DOB) and the PID controller operate on the slow timescale. With the Chinese time standard Under the following conditions, the sliding mode variable structure control (SMC) operates at a fast timescale. It operates under different disturbance scenarios and achieves dynamic balance through time-scale switching conditions.

[0020] Preferably, step S1 specifically includes:

[0021] Define path stability deviation as: ;in, This refers to the number of nodes (including warehousing, loading and unloading, handover, and emergency repair stations) that match the actual transportation of power materials with the planned routes. The total number of nodes in the planned route for power materials, with a value range of [value missing]. ;

[0022] Define the cumulative deviation of node disturbances as: ;in, For the first Delay time at each power supply node, This represents the total number of nodes contained in the path. The preset power material node delay threshold;

[0023] The degree to which the overall system operation deviates from steady state: ;in, The number of Key Performance Indicators (KPIs) For the first The actual value of each indicator This is the steady-state value of the index, and its range is [value missing]. ;

[0024] Reference steady state Constructed using the sliding window statistical method: ;in, The length of the sliding window. For the first The system state vector at time t.

[0025] Preferably, step S3 specifically includes:

[0026] The disturbance observer employs a third-order extended state observer, and its mathematical model is as follows:

[0027] ;

[0028] in:

[0029] For the actual output of the system The estimated value, To provide an estimate of the first derivative of the system output, This is an estimate of the equivalent disturbance, i.e., the disturbance value observed in real time. To output the estimated value rate of change over time The differential estimate rate of change over time Equivalent disturbance estimate Rate of change over time;

[0030] , and Let be the observer gain coefficient, satisfying , , ; This represents the observer bandwidth, with a value range based on real-world conditions.

[0031] This is the original control input of the system, i.e., the basic control quantity without feedforward compensation; For the system control input gain, specifically ,in, For the system static gain, The system time constant;

[0032] , ; This is the width of the linear interval of the nonlinear function, ranging from 0.001 to 0.01, used to avoid high-frequency chattering;

[0033] To replace traditional linear functions and improve observer performance, specifically:

[0034] When the error is greater than When using a power function, convergence is fast, and the error is less than [value missing]. Use linear terms to avoid chattering;

[0035] Preferably, the system nominal model uses a first-order inertial element to describe the dynamic characteristics of the power material transportation system: ;in, For the system's nominal model, The Laplace transform of the nominal model output. Laplace transform for control input;

[0036] Then by comparing the system nominal model Response to actual state The difference is used to estimate the equivalent disturbance in real time, which includes external environmental disturbances and internal model uncertainties. The combined effects of form;

[0037] The equivalent disturbance is applied to the control input of the power material transportation system using a feedforward compensation method. ;in, To compensate for the control input, This is the basic control input without feedforward compensation. This is the control input after feedforward compensation, which reduces the impact of disturbances on the system.

[0038] Preferably, step S4 specifically includes:

[0039] The PID controller with parameter self-tuning adopts a piecewise integral + parameter self-tuning strategy, and the control law is:

[0040] ;

[0041] in, ; To provide a unified error signal for the system; This represents the error component caused by the disturbance. For equivalent perturbation, For the general standard model;

[0042] Further,

[0043] The proportional coefficient self-tuning model is as follows: ; This is the initial scaling factor. This is the proportional coefficient attenuation factor;

[0044] The adaptive model for integral coefficients is as follows: ; These are the initial integration coefficients. The attenuation factor is the integral coefficient. , This is the integration threshold;

[0045] The piecewise integral function is: ; This refers to the width of the transition interval.

[0046] Differential filtering is: ; The time constant of the differential filter is... is the complex frequency variable in the Laplace transform, used to convert time-domain differential equations into complex frequency-domain algebraic equations.

[0047] Preferably, the sliding surface in step S5 is a nonlinear power sliding surface, expressed as:

[0048] ;

[0049] in, For the sliding surface state quantity, For error gain, For the gain of the rate of change of error, As a nonlinear factor, through , Optimize parameters, ω is the angular velocity of the sliding surface convergence.

[0050] Preferably, step S6 specifically includes:

[0051] When satisfied or At that time, the sliding mode variable structure control (SMC) is activated; among which, The threshold value for the sliding surface. The maximum effective disturbance intensity for DOB+PID collaborative control;

[0052] The sliding mode variable structure control employs adaptive gain switching and boundary layer chattering suppression, and the control law is as follows:

[0053] ;

[0054] in, For adaptive gain switching, Based on switching gain, For perturbation adaptive coefficients; It is a saturation function. Boundary layer thickness; It is the exponential approach rate coefficient for sluice membrane variable structure control, used to adjust the dynamic characteristics of the system converging towards the sluice membrane surface; As a reference error rate of change, This is the reference convergence coefficient.

[0055] Preferably, step S7 specifically includes:

[0056] Based on the control input after feedforward compensation PID controller with parameter self-tuning and sliding mode variable structure control Analysis yields comprehensive control output : ,in , To control the output weights;

[0057] The integrated control output mapping adopts a hybrid linear and nonlinear mapping model, including:

[0058] Path weight correction model: Further path weight correction is obtained. ;in, The original path weights, For the corrected weights, This is the weighting adjustment coefficient. For the first Perturbation estimates for each path;

[0059] Node avoidance factor adjustment formula: Further adjustments to the node avoidance factor were obtained. ;in, The original avoidance factor, For the adjusted factor, To avoid the adjustment coefficient of the factor, For the first Cumulative deviation of disturbances at each node;

[0060] Replanning triggers threshold update model: This further leads to the replanning trigger threshold update. ;in, As the initial threshold, The updated threshold Let be the threshold adjustment coefficient, and satisfy . , , These are the upper and lower limits of the threshold, respectively.

[0061] Preferably, step S8 specifically includes:

[0062] The timescale standard is defined as: slow timescale China Times fast time marker ;

[0063] The time scale switching condition is as follows: ;in, The disturbance threshold;

[0064] Closed-loop update iteration formula: ;in, To update the coefficients, The current runtime period;

[0065] The stability determination model is based on the Lyapunov function. ,in For the disturbance estimation error, satisfying , , This is a stability parameter to ensure the stability of the system index.

[0066] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0067] This invention constructs a hierarchical control architecture of "feedforward compensation-smooth adjustment-rigid constraint" by defining controllable state variables and reference steady state at the planning level, combining feedforward compensation with disturbance observers, improving the smooth adjustment of PID control, and rigid constraints of sliding mode variable structure control. It embeds a multi-dimensional mathematical model to achieve quantitative control and maps the control output to planning decision variables through a multi-timescale closed-loop update mechanism. Without interfering with the physical execution layer, it dynamically balances the system planning stability, disturbance resistance, and response efficiency, thus achieving a natural connection between control theory and power material dispatching, emergency repair path planning, and supply guarantee decision-making. Attached Figure Description

[0068] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0069] Figure 1 This is a flowchart of an anti-disturbance method for a power material transportation system based on feedforward compensation and robust control, as described in an embodiment of the present invention. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] Example 1: Combination Figure 1 A flowchart of the steps of an anti-disturbance method for a power material transportation system based on feedforward compensation and robust control provided by an embodiment of the present invention, the method comprising:

[0072] S1. Define the controllable state variables and reference steady state of the power material transportation system: obtain the state variables used for feedback control in the power material transportation system, construct the system state vector describing the deviation at the planning level, and determine the reference steady state of the system state vector through the sliding window statistical method or the initial solution of the planning.

[0073] The deviation includes path stability deviation, node disturbance cumulative deviation, and the degree to which the overall system operation deviates from steady state;

[0074] S2. Constructing the system error signal: Based on the difference between the current system state vector and the reference steady state, the three deviations—path stability deviation, cumulative node disturbance deviation, and the degree of overall system operation deviating from the steady state—are dynamically weighted to generate a unified system error signal. The error signal does not distinguish the source of the disturbance, providing a unified feedback interface for different control modules;

[0075] S3. Real-time Estimation and Feedforward Compensation of Unknown Disturbances: A third-order extended state observer is embedded in the system feedback channel as a disturbance observer. By comparing the difference between the nominal system model and the actual state response, the equivalent disturbance, including external environmental disturbances and internal model uncertainties, is estimated in real time. The combined effects of the forms are considered, and the equivalent disturbance is applied to the control input of the power material transportation system through feedforward compensation, so as to suppress the amplified impact of the disturbance on the timeliness of power material transportation and emergency repair guarantee in advance.

[0076] S4. Improved PID control for adjusting system deviation: Based on the system error signal and equivalent disturbance, the compensated error signal is confirmed, and a PID controller with parameter self-tuning is used to continuously and smoothly adjust the deviation of the power material transportation system; the proportional term suppresses the trend of deviation expansion, the integral term eliminates long-term accumulated deviation, and the derivative term suppresses system oscillation; the PID controller acts on the planning layer parameter update frequency and adjustment range;

[0077] S5. Constructing the sliding surface and defining the stability boundary: based on systematic errors. and its rate of change Constructing the sliding surface The sliding surface does not depend on an accurate system model and describes the dynamic trajectory of the power material transportation system converging from an unstable state to a stable state by setting switching rules.

[0078] S6. Sliding mode variable structure control to cope with strong disturbances: when the sliding surface state quantity or disturbance intensity When the effective range is exceeded, sliding mode variable structure control (SMC) is activated. The sliding mode variable structure control enhances the robust control effect by adaptively switching the control law, so that the system state converges quickly along the sliding surface.

[0079] S7. Mapping control output to planning decision variables: Based on the feedforward compensated control input, the parameterized self-tuning PID controller, and the sliding mode variable structure control analysis, the control output is synthesized. This is then transformed into executable decision adjustment quantities for the power material transportation planning layer through a mapping function. These decision adjustment quantities include path weight corrections. Node avoidance factor adjustment and replanning trigger threshold update ;

[0080] S8. Closed-loop update and multi-timescale collaborative operation: Steps S1-S7 are continuously executed in a closed loop, and state updates are achieved through iterative formulas. The disturbance observer (DOB) and the PID controller operate on the slow timescale. With the Chinese time standard Under the following conditions, the sliding mode variable structure control (SMC) operates at a fast timescale. It operates under different disturbance scenarios and achieves dynamic balance through time-scale switching conditions.

[0081] To better illustrate this, as an optional implementation method, in this embodiment, step S1 involves acquiring warehousing and distribution nodes based on the transportation of power materials (such as ultra-high voltage equipment and emergency power distribution materials). This is done by analyzing 100 power material circulation nodes in a certain region of China (including provincial central warehouses, municipal transfer warehouses, and remote substations). The region covers roughly from western to eastern China, including inland and coastal cities. Forty logistics nodes are selected from each region, and the number is increased to 100 nodes according to the algorithm's iteration process. This involves dynamically increasing the path planning based on the number of nodes, selecting some nodes for training, and gradually increasing to the maximum number of nodes within the entire power material distribution network. This forms a cross-regional multimodal transport network, which can comprehensively reflect the characteristics and needs of power grid material circulation in different regions, making the path planning more consistent with actual emergency repair and engineering conditions.

[0082] Preferably, the system parameters are set to the initial values ​​of the state variable weights: , , The reference steady state was obtained by referring to historical data from recent months. The DOB parameter is set as follows: , , , , , , The nominal model parameters are set as follows: , , PID parameters are set as follows: , , , , , , , The sliding surface parameters are set as follows: , , , SMC parameters are set as follows: , , , , , Multi-timescale period settings are as follows: , , .

[0083] Preferably, the system parameters are substituted into two transportation states. The first is a light to moderate disturbance scenario, which is a normal power supply situation, such as slight congestion on the power material transportation route and delays in normal warehousing and unloading. The system mainly uses DOB+PID control, does not trigger emergency repair rescheduling, smoothly corrects path and node deviations, and ensures stable supply timeliness without additional costs. The second is a strong disturbance scenario, which is a power emergency repair or extreme disaster scenario, such as typhoons causing road interruptions, port closures, and power grid line failures requiring emergency repair materials. When the disturbance intensity exceeds the limit, the system quickly switches to SMC fast timescale control, immediately adjusts the transportation route, avoids fault nodes, triggers emergency repair rescheduling, and converges to a stable state in a short time to ensure that power grid repair materials arrive on time.

[0084] Example 2: This example demonstrates a first type of light to moderate disturbance scenario, such as minor congestion on power supply transportation routes and delays in routine warehousing and loading / unloading; the specific implementation steps include:

[0085] S1: Define the controllable state variables and reference steady state of the power material transportation system: Select state variables that can be used for feedback control, and construct a system state vector describing the deviation at the planning level. The deviation includes path stability deviation. Node disturbance cumulative deviation And the degree to which the overall system operation deviates from steady state. Based on historical stable operating intervals or initial planning solutions, the system reference steady state is determined through statistical modeling. This serves as the control target for subsequent error calculations;

[0086] Preferably, the path stability deviation is defined as: ,in This refers to the number of nodes (including warehousing, loading and unloading, handover, and emergency repair stations) that match the actual transportation of power materials with the planned routes. The total number of nodes in the planned route for power materials, with a value range of [value missing]. ;

[0087] Define the cumulative deviation of node disturbances as: ,in For the first The delay time of each node, This represents the total number of nodes contained in the path. The preset delay threshold is used;

[0088] System status detection obtained , , State vector .

[0089] S2. Based on the difference between the current system state vector and the reference steady state, the three deviations—path stability deviation, node disturbance cumulative deviation, and the degree of overall system operation deviating from the steady state—are dynamically weighted to generate a unified system error signal. .

[0090] Preferably, the degree to which the overall system operation deviates from steady state is defined: ,in The number of Key Performance Indicators (KPIs) For the first The actual value of each indicator This is the steady-state value of the index, and its range is [value missing]. ;

[0091] Define reference steady state Constructed using the sliding window statistical method: ,in The length of the sliding window. For the first The system state vector at time t;

[0092] Constructing the system error signal: based on the system state vector at the current moment With reference steady state The differences are analyzed through weighted fusion to generate a unified system error signal. The error signal does not distinguish the source of the disturbance, providing a unified feedback interface for different control modules;

[0093] Dynamic weights , , Unified error signal .

[0094] S3. Embed a third-order extended state observer as a disturbance observer in the system feedback channel. By comparing the difference between the nominal system model and the actual state response, the equivalent disturbance, including external environmental disturbances and internal model uncertainties, is estimated in real time. The combined effects of the form are considered, and the equivalent disturbance is applied to the control input of the power material transportation system through feedforward compensation.

[0095] Preferably, real-time estimation and feedforward compensation of unknown disturbances: a disturbance observer (DOB) is embedded in the system feedback channel, based on the system nominal model. Response to actual state The difference, in terms of equivalent perturbation The system formally estimates the combined impact of external disturbances and model uncertainties in real time, and applies the equivalent disturbance to the power material transportation control input through feedforward compensation to suppress disturbance amplification in advance. The disturbance observer (DOB) used is an extended state observer, and its mathematical model is as follows:

[0096] ;

[0097] in:

[0098] For the actual output of the system The estimated value, To provide an estimate of the first derivative of the system output, This is an estimate of the equivalent disturbance, i.e., the disturbance value observed in real time. To output the estimated value rate of change over time The differential estimate rate of change over time Equivalent disturbance estimate Rate of change over time;

[0099] , and Let be the observer gain coefficient, satisfying , , ; This represents the observer bandwidth, with a value range based on real-world conditions. For the system control input gain, specifically ,in For the system static gain, The system time constant; This refers to the system's original control input and the basic control quantity without feedforward compensation.

[0100] , , , To replace traditional linear functions and improve observer performance, specifically: When the error is greater than When using a power function, convergence is fast, and the error is less than [value missing]. Use linear terms to avoid chattering; among them, This is the width of the linear interval of the nonlinear function, ranging from 0.001 to 0.01, used to avoid high-frequency chattering.

[0101] Preferably, the system nominal model The dynamic characteristics of the power material transportation system are described using a first-order inertial element: ;in, The Laplace transform of the nominal model output. Laplace transform for control input;

[0102] For the compensated control input, where, This is the basic control input without feedforward compensation. The control input is after feedforward compensation, which can reduce the impact of disturbances on the system;

[0103] DOB estimates equivalent perturbation After feedforward compensation , .

[0104] S4. Based on the system error signal and equivalent disturbance, confirm the compensated error signal, and use a PID controller with parameter self-tuning to continuously and smoothly adjust the deviation of the power material transportation system.

[0105] Preferably, the deviation of the PID control system is improved based on the error signal after compensation in step S3. A proportional-integral-derivative (PID) controller with parameter self-tuning is used to continuously and smoothly adjust the deviation of the power material transportation system. The proportional term suppresses the trend of deviation expansion, the integral term eliminates long-term accumulated deviation, and the derivative term suppresses system oscillation. The PID controller acts on the update frequency and adjustment range of the planning layer parameters.

[0106] The improved PID controller described above employs a piecewise integral + parameter self-tuning strategy, and its control law is as follows:

[0107] ;

[0108] in Original systematic error The unified error signal obtained by weighted fusion in step S2; This represents the error component caused by the disturbance. The equivalent perturbation estimated by DOB in step S3. It is a standard model for the system.

[0109] The proportional coefficient self-tuning model is as follows: , This is the initial scaling factor. This is the proportional coefficient attenuation factor;

[0110] The adaptive model for integral coefficients is as follows: , , The integration threshold is... These are the initial integration coefficients. This is the attenuation factor for the integral coefficient;

[0111] Among them, piecewise integral function ; This represents the width of the transition interval.

[0112] Among them, differential term filtering: , The time constant of the differential filter is... is the complex frequency variable in the Laplace transform, used to convert time-domain differential equations into complex frequency-domain algebraic equations.

[0113] PID parameter self-tuning: , (because ), integral function Differential term , .

[0114] S5. Constructing the sliding surface and defining the stability boundary: based on systematic errors. and its rate of change Constructing the sliding surface The sliding surface does not depend on an accurate system model and describes the dynamic trajectory of the power material transportation system converging from an unstable state to a stable state by setting switching rules.

[0115] S6. Sliding mode variable structure control to cope with strong disturbances: when the sliding surface state quantity or disturbance intensity When the effective range is exceeded, sliding mode variable structure control (SMC) is activated. The sliding mode variable structure control strengthens the robust control effect by adaptively switching the control law, so that the system state converges quickly along the sliding surface.

[0116] Preferably, the sliding mode variable structure control employs adaptive gain switching + boundary layer chattering suppression, and the control law is: ;

[0117] Wherein: Adaptive switching gain: , Based on switching gain, For the perturbation adaptive coefficient; saturation function: Boundary layer thickness; Reference error rate of change: , The reference convergence coefficients are used; the SMC start-up condition is: or ,in The threshold value for the sliding surface. This represents the maximum effective disturbance intensity for DOB+PID collaborative control. It is the exponential approach rate coefficient for synovial variable structure control, used to adjust the dynamic characteristics of the system converging towards the synovial surface.

[0118] Sliding surface Actual calculation: , , , ;but Therefore, SMC is not enabled.

[0119] S7. Mapping control output to planning decision variables: Based on the feedforward compensated control input, the parameterized self-tuning PID controller, and the sliding mode variable structure control analysis, the control output is synthesized. This is then transformed into executable decision adjustment quantities for the power material transportation planning layer through a mapping function. These decision adjustment quantities include path weight corrections. Node avoidance factor adjustment and replanning trigger threshold update ;

[0120] The path weight correction model is as follows: , The original path weights, For the corrected weights, This is the weighting adjustment coefficient. For the first Perturbation estimates for each path;

[0121] The formula for adjusting the node avoidance factor is: , The original avoidance factor, For the adjusted factor, To avoid the adjustment coefficient of the factor, For the first Cumulative deviation of disturbances at each node;

[0122] The replanning trigger threshold update model is as follows: , As the initial threshold, The updated threshold Let be the threshold adjustment coefficient, and satisfy . , , These are the upper and lower limits of the threshold, respectively;

[0123] Integrated control output: ,in , To control the output weights;

[0124] Therefore, the comprehensive control output is obtained. (in steady state) (term convergence)

[0125] Path weight correction: Highways .

[0126] S8. Closed-loop update and multi-timescale collaborative operation: Steps S1-S7 are continuously executed in a closed loop, and state updates are achieved through iterative formulas. The disturbance observer (DOB) and the PID controller operate on the slow timescale. With the Chinese time standard Under the following conditions, the sliding mode variable structure control (SMC) operates at a fast timescale. It operates under different disturbance scenarios and achieves dynamic balance through time-scale switching conditions.

[0127] The time scale definition standard is: slow time scale. China Times fast time marker ;

[0128] The time scale switching condition is: ,in The disturbance threshold;

[0129] Closed-loop update iteration formula: ,in To update the coefficients, The current runtime period;

[0130] The stability determination model is based on the Lyapunov function. ( (for the disturbance estimation error), satisfying , , These are stability parameters to ensure the stability of the system index;

[0131] Updated in a closed loop according to the medium time scale (3 minutes). The system gradually returned to a steady state without triggering replanning.

[0132] Example 3: This example demonstrates a scenario with strong disturbances, such as a typhoon causing road closures, port shutdowns, or power grid line failures requiring emergency repair materials. When the disturbance intensity exceeds limits, the SMC fast timescale control is quickly switched on to immediately adjust transport routes, avoid faulty nodes, and trigger emergency repair rescheduling. The system converges to a stable state within a short time, ensuring the timely arrival of power grid repair materials. The specific implementation process is as follows:

[0133] S1: Define the controllable state variables and reference steady state of the power material transportation system: Select state variables that can be used for feedback control, and construct a system state vector describing the deviation at the planning level. The deviation includes path stability deviation. Node disturbance cumulative deviation And the degree to which the overall system operation deviates from steady state. Based on historical stable operating intervals or initial planning solutions, the system reference steady state is determined through statistical modeling. This serves as the control target for subsequent error calculations;

[0134] Preferably, the path stability deviation is defined as: ,in This refers to the number of nodes (including warehousing, loading and unloading, handover, and emergency repair stations) that match the actual transportation of power materials with the planned routes. The total number of nodes in the planned route for power materials, with a value range of [value missing]. ;

[0135] Define the cumulative deviation of node disturbances as: ,in For the first Delay time at each node of power supply transportation, This represents the total number of nodes contained in the path. The preset delay threshold is used;

[0136] System status detection obtained , , State vector ;

[0137] Define the degree to which the overall system operation deviates from steady state: ,in The number of Key Performance Indicators (KPIs) For the first The actual value of each indicator This is the steady-state value of the index, and its range is [value missing]. .

[0138] S2: Define the reference steady state Constructed using the sliding window statistical method: ,in The length of the sliding window. For the first The system state vector at time t.

[0139] Constructing the system error signal: based on the system state vector at the current moment With reference steady state The differences are analyzed through weighted fusion to generate a unified system error signal. The error signal does not distinguish the source of the disturbance, providing a unified feedback interface for different control modules;

[0140] Dynamic weights , , Unified error signal ;

[0141] S3: Real-time estimation and feedforward compensation of unknown disturbances: An observation of disturbances (DOB) is embedded in the system feedback channel, based on the system nominal model. Response to actual state The difference, in terms of equivalent perturbation The system formally estimates the combined impact of external disturbances and model uncertainties in real time. Through feedforward compensation, the equivalent disturbance is applied to the power material transportation control input, thus suppressing the amplified impact of disturbances on the timeliness of power material transportation and emergency repair support. The disturbance observer (DOB) used is an extended state observer, and its mathematical model is as follows: ;

[0142] in:

[0143] For the actual output of the system The estimated value, To provide an estimate of the first derivative of the system output, This is an estimate of the equivalent disturbance, i.e., the disturbance value observed in real time. To output the estimated value rate of change over time The differential estimate rate of change over time Equivalent disturbance estimate Rate of change over time;

[0144] , and Let be the observer gain coefficient, satisfying , , ; This represents the observer bandwidth, with a value range based on real-world conditions. For the system control input gain, specifically ,in For the system static gain, The system time constant; This refers to the system's original control input and the basic control quantity without feedforward compensation.

[0145] , , , To replace traditional linear functions and improve observer performance, specifically: When the error is greater than When using a power function, convergence is fast, and the error is less than [value missing]. Use linear terms to avoid chattering; among them, This is the width of the linear interval of the nonlinear function, ranging from 0.001 to 0.01, used to avoid high-frequency chattering.

[0146] Preferably, the system nominal model uses a first-order inertial element to describe the dynamic characteristics of the power material transportation system;

[0147] ;in, The Laplace transform of the nominal model output. Laplace transform for control input;

[0148] For the compensated control input, where This is the basic control input without feedforward compensation. The control input is after feedforward compensation, which can reduce the impact of disturbances on the system;

[0149] DOB estimates equivalent perturbation After feedforward compensation , .

[0150] S4: Improve the deviation of the PID control system: based on the error signal after compensation in step S3. A proportional-integral-derivative (PID) controller with parameter self-tuning is used to continuously and smoothly adjust the deviation of the power material transportation system. The proportional term suppresses the tendency for deviation to increase, the integral term eliminates long-term accumulated deviation, and the derivative term suppresses system oscillation. The PID controller operates on the parameter update frequency and adjustment range of the power material transportation planning layer.

[0151] The improved PID controller described above employs a piecewise integral + parameter self-tuning strategy, and its control law is as follows:

[0152] ;

[0153] in Original systematic error The unified error signal obtained by weighted fusion in step S2; This represents the error component caused by the disturbance. The equivalent perturbation estimated by DOB in step S3. For the general standard model;

[0154] The proportional coefficient self-tuning model is as follows: , This is the initial scaling factor. This is the proportional coefficient attenuation factor;

[0155] The adaptive model for the integral coefficients is as follows: , , The integration threshold is... These are the initial integration coefficients. This is the attenuation factor for the integral coefficient;

[0156] Where the piecewise integral function is: ; This refers to the width of the transition interval.

[0157] Among them, differential term filtering: , The time constant of the differential filter is... is the complex frequency variable in the Laplace transform, used to convert time-domain differential equations into complex frequency-domain algebraic equations.

[0158] PID parameter self-tuning: , (because ), integral function (but , , Therefore ), ;

[0159] S5: Constructing the sliding surface to characterize the stability boundary: based on systematic errors and its rate of change Constructing the sliding surface The sliding surface does not depend on an accurate system model and describes the dynamic trajectory of the power material transportation system converging from an unstable state to a stable state by setting switching rules.

[0160] S6: Sliding mode variable structure control for handling strong disturbances: When the error rate of change is detected... or disturbance intensity When the effective range of the coordinated control in steps S3 and S4 is exceeded, sliding mode variable structure control (SMC) is activated. The robust control effect is enhanced by adaptive switching control law, so that the system state converges quickly along the sliding surface.

[0161] The sliding mode variable structure control employs adaptive gain switching and boundary layer chattering suppression, with the control law being:

[0162] ;

[0163] Wherein: Adaptive switching gain: , Based on switching gain, For perturbation adaptive coefficients; saturation function: Boundary layer thickness; Reference error rate of change: , The reference convergence coefficients are used; the SMC start-up condition is: or ,in The threshold value for the sliding surface. This represents the maximum effective disturbance intensity for DOB+PID collaborative control. It is the exponential approach rate coefficient for sluice membrane variable structure control, used to adjust the dynamic characteristics of the system converging towards the sluice membrane surface;

[0164] Sliding surface Enable SMC;

[0165] SMC control output , , , ;

[0166] S7: Mapping control output to planning decision variables: Mapping the integrated control output processed by steps S3, S4, and S6... The decision adjustment is transformed into an executable decision adjustment quantity at the power material transportation planning layer through a mapping function. The decision adjustment quantity includes path weight correction. Node avoidance factor adjustment and replanning trigger threshold update ;

[0167] The path weight correction model is as follows: , The original path weights, For the corrected weights, This is the weighting adjustment coefficient. For the first Perturbation estimates for each path;

[0168] The formula for adjusting the node avoidance factor is: , The original avoidance factor, For the adjusted factor, To avoid the adjustment coefficient of the factor, For the first Cumulative deviation of disturbances at each node;

[0169] The replanning trigger threshold update model is as follows: , As the initial threshold, The updated threshold Let be the threshold adjustment coefficient, and satisfy . , , These are the upper and lower limits of the threshold, respectively;

[0170] Integrated control output: ,in , To control the output weights;

[0171] Integrated control output Path weight correction: maritime transport ,railway ,highway Node avoidance factor (Port Node); Replanning Trigger Threshold This triggers a replanning process.

[0172] S8: Closed-loop update and multi-timescale collaborative operation: Steps S1-S7 are continuously executed in a closed-loop manner, and the state is updated through iterative formulas. Among them, the DOB in step S3 and the PID control in step S4 are performed on the slow timescale. With the Chinese time standard Under the following operation, the SMC in step S6 is at the fast timescale The system operates under different disturbance scenarios, achieving dynamic equilibrium through time-scale switching conditions.

[0173] The time scale definition standard is: slow time scale. China Times fast time marker ;

[0174] The time scale switching condition is: ,in The disturbance threshold;

[0175] Closed-loop update iteration formula: ,in To update the coefficients, The current runtime period;

[0176] The stability determination model is based on the Lyapunov function. ( (for the disturbance estimation error), satisfying , , This is a stability parameter to ensure the stability of the system index.

[0177] Updated in a closed loop using fast timescales (0.5 minutes). The system state converges rapidly along the sliding surface and returns to stability within 10 minutes.

[0178] Finally, based on Examples 2 and 3, it was found that the power material transportation system did not experience frequent replanning under light to moderate disturbances, with a steady-state recovery time ≤15 minutes and an error convergence rate ≥0.02 / min. Under strong disturbances, it quickly switched control strategies, with a replanning response time ≤1 minute, a steady-state recovery time ≤10 minutes, transportation timeliness fluctuations controlled within ±8%, and cost increases ≤10%, significantly outperforming traditional control methods (steady-state recovery time ≥30 minutes, timeliness fluctuations ±15%, and cost increases ≥20%). With the quantitative support of the mathematical model, the system control accuracy was improved by more than 40%, and its robustness against disturbances was significantly enhanced.

[0179] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A disturbance rejection method for a power material transportation system based on feedforward compensation and robust control, characterized in that, Includes the following steps: S1. Obtain the state variables used for feedback control in the power material transportation system, construct a system state vector describing the deviation at the planning level, and determine the reference steady state of the system state vector through the sliding window statistical method or the initial solution of the planning; wherein, the deviation includes path stability deviation, node disturbance cumulative deviation, and the degree of deviation of the overall system operation from the steady state; S2. Based on the difference between the current system state vector and the reference steady state, the three deviations—path stability deviation, node disturbance cumulative deviation, and the degree of overall system operation deviating from the steady state—are dynamically weighted to generate a unified system error signal. ; S3. Embed a third-order extended state observer as a disturbance observer in the system feedback channel. By comparing the difference between the nominal system model and the actual state response, the equivalent disturbance, including external environmental disturbances and internal model uncertainties, is estimated in real time. The combined effects of the form, and the equivalent disturbance is compensated through feedforward. It acts on the control input terminal of the power material transportation system; S4. Based on the system error signal and equivalent disturbance, confirm the compensated error signal, and use a PID controller with parameter self-tuning to continuously and smoothly adjust the deviation of the power material transportation system. S5, Based on system error signal and its rate of change Constructing the sliding surface The sliding surface To describe the dynamic trajectory of the power material transportation system converging from an unstable state to a stable state by setting switching rules; S6. Take the absolute values ​​of the sliding surface and the equivalent disturbance respectively. If the sliding surface state quantity or equivalent disturbance strength When the effective range is exceeded, sliding mode variable structure control is activated. The sliding mode variable structure control enhances the robust control effect by adaptively switching the control law, so that the system state converges quickly along the sliding surface. S7. Based on the feedforward compensated control input, the parameter-self-tuning PID controller, and the sliding mode variable structure control analysis, the comprehensive control output is determined. This is then transformed into executable decision adjustment quantities for the power material transportation planning layer through a mapping function. These decision adjustment quantities include path weight corrections. Node avoidance factor adjustment and replanning trigger threshold update ; S8. Continuously execute steps S1-S7 in a closed-loop manner, updating the state through iterative formulas. The disturbance observer (DOB) and the PID controller operate on a slow timescale. With the Chinese time standard Under the following conditions, the sliding mode variable structure control (SMC) operates at a fast timescale. It operates under different disturbance scenarios and achieves dynamic balance through time-scale switching conditions.

2. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, Step S1 specifically includes: Define path stability deviation as: ;in, This represents the number of nodes matching the actual transportation of power supplies with the planned routes. The total number of nodes in the planned route for power materials, with a value range of [value missing]. ; Define the cumulative deviation of node disturbances as: ;in, For the first Delay time at each power supply node, This represents the total number of nodes contained in the path. The preset threshold for power material node delay; The degree to which the overall system operation deviates from steady state: ;in, The number of key performance indicators. For the first The actual value of each indicator This is the steady-state value of the index, and its range is [value missing]. ; Reference steady state Constructed using the sliding window statistical method: ;in, The length of the sliding window. For the first The system state vector at time t.

3. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, The disturbance observer mentioned in step S3 adopts a third-order extended state observer, whose mathematical model is as follows: ； in: For the actual output of the system The estimated value, To provide an estimate of the first derivative of the system output, This is an estimate of the equivalent disturbance, i.e., the disturbance value observed in real time. To output the estimated value rate of change over time The differential estimate rate of change over time Equivalent disturbance estimate Rate of change over time; , and Let be the observer gain coefficient, satisfying , , ; For observer bandwidth; This is the original control input of the system, i.e., the basic control quantity without feedforward compensation; For the system control input gain, specifically ,in, For the system static gain, The system time constant; , ; The width of the linear interval of the nonlinear function; To replace traditional linear functions and improve observer performance, specifically: .

4. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, The nominal system model described in step S3 uses a first-order inertial element to describe the dynamic characteristics of the power material transportation system. The specific formula is as follows: ； in, For the system's nominal model, The Laplace transform of the nominal model output. Laplace transform for control input; Then by comparing the system nominal model Response to actual state The difference is used to estimate the equivalent disturbance in real time, which includes external environmental disturbances and internal model uncertainties. The combined effects of form.

5. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, The equivalent disturbance described in step S3 is compensated by feedforward. The formula applied to the control input of the power material transportation system is as follows: ； in, To compensate for the control input, This is the basic control input without feedforward compensation. This is the control input after feedforward compensation.

6. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, Step S4 specifically includes: The PID controller with parameter self-tuning adopts a piecewise integral + parameter self-tuning strategy, and the control law is: ； in, ; To provide a unified error signal for the system; This represents the error component caused by the disturbance. For equivalent perturbation, For the general standard model; Further, The proportional coefficient self-tuning model is as follows: ; This is the initial scaling factor. This is the proportional coefficient attenuation factor; The adaptive model for integral coefficients is as follows: ; These are the initial integration coefficients. The attenuation factor is the integral coefficient. , This is the integration threshold; The piecewise integral function is: ; This refers to the width of the transition interval. Differential filtering is: ; The time constant of the differential filter is... is the complex frequency variable in the Laplace transform.

7. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, The sliding surface mentioned in step S5 is a nonlinear power sliding surface, and its expression is: ； in, For the sliding surface state quantity, For error gain, For the gain of the rate of change of error, As a nonlinear factor, through , Optimize parameters, ω is the angular velocity of the sliding surface convergence.

8. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, Step S6 specifically includes: When satisfied or At that time, the sliding mode variable structure control (SMC) is activated; among which, The threshold value for the sliding surface. The maximum effective disturbance intensity for DOB+PID collaborative control; The sliding mode variable structure control employs adaptive gain switching and boundary layer chattering suppression, and the control law is as follows: ； in, For adaptive gain switching, Based on switching gain, For perturbation adaptive coefficients; It is a saturation function. Boundary layer thickness; The exponential approach rate coefficient for synovial variable structure control; As a reference error rate of change, This is the reference convergence coefficient.

9. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, Step S7 specifically includes: Based on the control input after feedforward compensation PID controller with parameter self-tuning and sliding mode variable structure control Analysis yields comprehensive control output : ,in , To control the output weights; The integrated control output mapping adopts a linear + nonlinear hybrid mapping model, including: Path weight correction model: Further path weight correction is obtained. ;in, The original path weights, For the corrected weights, This is the weighting adjustment coefficient. For the first Perturbation estimates for each path; Node avoidance factor adjustment formula: Further adjustments to the node avoidance factor were obtained. ;in, The original avoidance factor, For the adjusted factor, To avoid the adjustment coefficient of the factor, For the first Cumulative deviation of disturbances at each node; Replanning triggers threshold update model: This further leads to the replanning trigger threshold update. ;in, As the initial threshold, The updated threshold Let be the threshold adjustment coefficient, and satisfy . , , These are the upper and lower limits of the threshold, respectively.

10. The disturbance rejection method for a power material transportation system based on feedforward compensation and robust control as described in claim 1, characterized in that, Step S8 specifically includes: The timescale standard is defined as: slow timescale China Times fast time marker ; The time scale switching conditions are as follows: ;in, The disturbance threshold; Closed-loop update iteration formula: ;in, To update the coefficients, The current runtime period; The stability determination model is based on the Lyapunov function. ,in For the disturbance estimation error, satisfying , , This is a stability parameter to ensure the stability of the system index.

Images