New energy truck charging path co-scheduling optimization system

Abstract

The invention discloses a new energy truck charging path collaborative scheduling optimization system, and relates to the technical field of logistics transportation scheduling, and the system comprises a data fusion processing unit, an AI model calculation unit, and a scheduling decision unit. The data fusion processing unit is used for acquiring vehicle position, residual electric quantity, charging station state and power grid load data and performing integration processing on the data; the AI model calculation unit is used for constructing a dynamic demand prediction model, a multi-objective optimization scheduling model and a power grid-charging station-truck cooperative control model, and calculating truck charging dynamic demand information, optimization information and a control path based on each model; and the scheduling decision unit is used for generating an optimal charging path and a scheduling strategy, receiving feedback results after the vehicle-mounted navigation system and the charging pile control system execute the scheduling instruction, and carrying out visual monitoring on the charging information and the charging station data of each truck. The power grid load is more stable; and meanwhile, the transportation cost is reduced.

Description

Technical Field

[0001] This application relates to the field of logistics transportation scheduling technology, specifically to a collaborative scheduling optimization system for charging routes of new energy freight vehicles. Background Technology

[0002] With the widespread application of new energy trucks in the logistics industry, problems such as unreasonable charging infrastructure layout, low charging service efficiency, long queuing times, insufficient energy dispatch coordination, and a disconnect between route planning and charging demand have become bottlenecks restricting the sustainable development of the industry. From a spatial perspective, existing charging station site selection models are mostly based on static demand forecasting, failing to effectively integrate the dynamic spatiotemporal characteristics of logistics transportation. This results in a severe shortage of charging piles in high-frequency use areas such as transportation hubs and logistics parks, while equipment utilization rates in suburban and remote areas remain low for a long time. In the field of route planning, traditional algorithms do not consider key parameters such as the vehicle's real-time remaining battery power, charging station load status, and charging time, leading to a high risk of energy depletion during vehicle operation, affecting the timeliness and safety of logistics transportation. At the energy dispatch level, a collaborative optimization mechanism between the power grid, charging stations, and truck clusters has not yet been established. Existing dispatch strategies cannot achieve dynamic balance between power supply and demand, exacerbating peak-valley load fluctuations in the power grid and increasing system operating costs and safety hazards. Summary of the Invention

[0003] In view of this, this application provides a new energy truck charging route collaborative scheduling optimization system, which solves the technical problems existing in the field of new energy truck charging, such as long charging time, some charging scheduling systems only statically allocate charging piles without collaborative optimization with the route, unreasonable route planning, uneven utilization of charging piles, and large fluctuations in grid load.

[0004] To achieve the above objectives, the present invention provides the following technical solution: This invention mainly consists of a collaborative scheduling and optimization system for charging routes of new energy freight vehicles. The system constructs a three-layer architecture comprising a data fusion processing unit, an AI model calculation unit, and a scheduling decision unit, aiming to achieve efficient collaborative management of charging resources for new energy freight vehicles. The data fusion processing unit, as the system's perception layer, collects multi-source heterogeneous data in real time through vehicle terminals, charging pile sensors, road network monitoring systems, and power grid SCADA (Supervisory and Data Acquisition) systems. This data includes key information such as vehicle location, remaining battery power, charging station operating status, and power grid load. The collected data, after standardization and integration, is stored in a basic data pool, providing fundamental data support for subsequent intelligent analysis. The AI ​​model calculation unit is connected to the data fusion processing unit and constructs an algorithm matrix including a dynamic demand prediction model, a multi-objective optimization scheduling model, and a power grid-charging station-freight vehicle collaborative control model. Through model calculation, it achieves accurate prediction of dynamic demand for freight vehicle charging, global optimization of charging schemes, and collaborative regulation of the energy system. The scheduling decision unit acts as the execution center, generating the optimal charging path and scheduling strategy based on the output results of the AI ​​model. It also achieves closed-loop management of scheduling instructions through two-way communication with the vehicle navigation system and the charging pile control system, and realizes real-time monitoring of the entire system's operating status through a visual monitoring platform.

[0005] Furthermore, the data fusion processing unit consists of a data preprocessing module and a multi-source data fusion module, forming a complete data processing chain; The data preprocessing module interacts with the traffic management platform, power grid information system, and vehicle terminals via API interfaces. It is used to obtain the vehicle location, remaining battery power, connection status, output power, fault codes, substation load data, and real-time congestion index extracted from the traffic management platform API for each truck in the system. The module performs outlier cleaning, spatiotemporal alignment, and data compression preprocessing on the data. Outlier cleaning is based on the 3σ principle to remove data with sudden changes in SOC, such as fluctuations greater than 20% within 10 seconds. Spatiotemporal alignment unifies heterogeneous data into a spatiotemporal coordinate system composed of WGS-84 geographic coordinates and UTC timestamps. Data compression uses Delta encoding to reduce the storage of redundant trajectory points.

[0006] The multi-source data fusion module is used to extract features from the data, acquiring vehicle SOC information, current status information of charging piles, road network congestion index, electricity price time period, and grid load, and fusing them to generate a spatiotemporal feature matrix. Specifically, it uses the Kalman filter algorithm to fuse multi-sensor data to improve the accuracy of SOC estimation, performs one-hot encoding on charging pile status data to construct standardized feature vectors, extracts charging pile status (power, availability, fault codes), and calculates dynamic congestion information by combining real-time vehicle speed and historical traffic flow using a historical similarity matching method, achieving accurate prediction of the dynamic traffic congestion index. For grid load, it uses wavelet decomposition to perform frequency domain analysis on grid load data, extracting the fundamental frequency component to reflect the basic trend of grid load changes. Finally, it generates a feature matrix containing spatiotemporal dimensions, providing high-quality input data for subsequent model calculations.

[0007] Furthermore, the AI ​​model computing unit includes a demand forecasting module, an optimization scheduling module, and a collaborative control and adjustment module; The demand forecasting module is used to build a dynamic demand forecasting model based on the Long Short-Term Memory (LSTM) network. The dynamic demand forecasting model includes an input layer, a bidirectional LSTM layer, a spatiotemporal attention layer, a fully connected layer, and an output layer. The bidirectional LSTM layer captures the time-series features of charging demand, and the spatiotemporal attention mechanism enhances the model's ability to perceive regional differences in charging demand. Historical charging data, traffic flow, weather conditions, and electricity pricing policies are used as input features to the dynamic demand forecasting model to make spatiotemporal predictions of charging demand. The output is the probability distribution of charging demand for each charging station within a preset time period in the future, thereby improving the model's generalization ability.

[0008] The input dimensions of the dynamic demand forecasting model include historical charging parameters, temperature, and electricity price. An attention mechanism is used to calculate the regional charging demand weights Wi. The model training uses quantile loss to optimize long-tail distribution prediction and generates virtual charging scenarios under extreme weather conditions through a GAN network to augment training data.

[0009] The optimization scheduling module is used to construct a multi-objective optimization model with the goals of minimizing transportation costs, maximizing the utilization rate of charging facilities, and minimizing grid load fluctuations. The particle swarm optimization algorithm is used to solve the multi-objective optimization model to generate the optimal charging route for each truck. The transportation cost includes travel time and charging cost.

[0010] The coordinated control and adjustment module is used to implement dynamic adjustment strategies based on coordinated control conditions, including real-time electricity price incentives, charging power adjustment, and truck route adjustment. Real-time electricity price incentives include dynamically adjusting charging prices based on grid load conditions: lowering prices during off-peak hours to attract trucks to charge, and raising prices during peak hours to suppress charging demand. For example, when the grid load exceeds a preset threshold, the charging price is increased by 15% to guide peak-shaving. Charging power adjustment includes charging stations adjusting their output power in real time according to grid load. For example, if multiple trucks are connected to the same transformer simultaneously, the power is reduced to 80kW. Truck route adjustment includes automatically recommending alternative charging routes for trucks when the grid load is too high or the charging queue time is too long, avoiding concentrated charging. For example, if the charging queue time exceeds 30 minutes, the truck is redirected to an adjacent station.

[0011] Furthermore, the objective function of the multi-objective optimization model is: Where Ci is the transportation cost of the i-th truck, and Uj is the utilization rate of the j-th charging station. Let be the standard deviation of the grid load in time period t. , The weighting coefficients are used to define the constraints of the objective function: the remaining battery power of the vehicle is not lower than the safety threshold, the power of the charging station does not exceed the rated capacity, and the grid load does not exceed the peak limit.

[0012] Furthermore, the AI ​​model computing unit also includes a dynamic conflict resolution module. This module monitors charging pile competition in real time. If multiple vehicles approach the same charging pile, they are rescheduled according to their urgency. The urgency assessment includes SOC margin and task priority. The urgency assessment model is as follows: ,in, and These are the weighting coefficients. Remaining time for the task A deadline is set, and a virtual charging queue mechanism is introduced to plan temporary tasks for waiting vehicles.

[0013] Furthermore, the scheduling decision unit includes a charging path coordination module, an intelligent scheduling module, and an execution feedback module; The charging path coordination module receives the optimal charging path information for each truck and dynamically adjusts the single charging amount based on subsequent road conditions to optimize segmented charging.

[0014] The specific steps for optimizing segmented charging include: Input the vehicle's initial SOC (State of Charge, current battery percentage) and the planned route (containing multiple route segments, each with a start point, end point, distance, and estimated energy consumption information); set the current battery level as the initial SOC and create an empty charging plan list; then iterate through each route segment: a. Calculate the energy consumption required to pass through the current segment, which can be estimated using the segment distance, vehicle energy consumption model (e.g., kWh / km), and environmental factors (e.g., gradient, air conditioning usage); b. Predict the remaining battery power after passing through the segment: soc_after_segment = current_soc - (required_energy / battery) c. Calculate the remaining battery capacity after passing through the current segment by multiplying it by 100% and converting the percentage of battery capacity into actual battery capacity (kWh); d. Check if the remaining battery capacity after passing through the current segment is lower than the safety threshold or lower than the battery capacity required to reach the next charging point (or destination) plus a redundancy, such as 10%. If so, charging should be arranged at the end of the current segment, i.e., the next node. When determining the charging amount, charge until the battery capacity is sufficient to meet the requirements of the subsequent segments to reach the next charging point (or destination) plus a safety redundancy, but not exceeding 80% of the battery capacity. Avoid fully charging to save time, as the battery charging curve slows down after 80%; e. Update the current battery capacity: if charging is done at this node, the current battery capacity becomes the charged battery capacity; if not, the current battery capacity is soc_after_segment; finally, continue traversing the next path segment until the entire path is completed, and output a charging plan list, which includes charging location and charging amount.

[0015] The intelligent scheduling module is connected to the charging path coordination module. The intelligent scheduling module is used to send charging station and charging quantity recommendation information to the vehicle navigation system of the vehicle terminal based on the path optimization results, and to send charging adjustment parameters to the charging pile control system.

[0016] The execution feedback module is used to receive the execution results after the charging pile control system executes the scheduling instructions, and to perform visual monitoring of the charging information of each truck and the charging station data.

[0017] Furthermore, the scheduling decision unit also includes a user interaction module. The user interaction module is used to obtain the identity information of the management personnel, compare the obtained identity information with the personnel verification information in the basic database to determine whether they are legitimate personnel, and assign corresponding operation permissions according to role information. It is also used to obtain the user's query instructions and parameter editing instructions.

[0018] As can be seen from the above technical solution, the advantages of the present invention are: This invention solves the multi-objective collaborative optimization problem in the planning and scheduling of charging routes for new energy freight vehicles. By constructing a dynamic demand prediction model, a multi-objective optimization scheduling model, and a real-time collaborative control mechanism, it achieves improved utilization of charging facilities, stable grid load, and minimized transportation costs. Through dynamic demand prediction and multi-objective optimization scheduling, the utilization rate of charging facilities is improved. The collaborative control mechanism reduces grid load fluctuations, reduces wind and solar curtailment rates, and optimizes route planning and charging strategies, thereby reducing the transportation cost per vehicle by more than 15%. Furthermore, the real-time collaborative control mechanism effectively responds to sudden changes in demand and improves system stability. Attached Figure Description

[0019] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0020] Figure 1 This is a schematic diagram of the composition structure of the new energy truck charging route collaborative scheduling optimization system of this application.

[0021] Figure 2 This is a schematic diagram of the data fusion processing unit in this embodiment.

[0022] Figure 3 This is a schematic diagram of the segmented charging optimization process in this embodiment.

[0023] Figure 4 This is a schematic diagram illustrating the steps of the collaborative scheduling optimization method for charging routes of new energy freight vehicles in this embodiment. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and their descriptions are used to explain this application, but are not intended to limit it.

[0025] New energy trucks face limitations in range and long charging times. Blindly charging leads to low transportation efficiency. Uneven distribution of charging stations and a lack of dynamic scheduling mechanisms cause severe congestion during peak hours. Existing route planning systems do not integrate real-time charging demand, resulting in conflicts between suboptimal routes and charging needs. Traditional navigation systems only optimize route distance / time, neglecting charging constraints. Some charging scheduling systems statically allocate charging stations, lacking collaborative optimization with routes. Furthermore, they fail to address resource contention issues in multi-vehicle collaborative scheduling. Figures 1 to 4 This embodiment provides a collaborative scheduling optimization system for charging routes of new energy freight vehicles. By constructing a dynamic demand prediction model, a multi-objective optimization scheduling model, and a collaborative control mechanism between the power grid, charging stations, and freight vehicles, it achieves improved utilization of charging facilities, stable power grid load, and minimized transportation costs. Figure 1As shown, the platform includes a data fusion processing unit, an AI model calculation unit, and a scheduling decision unit. The data fusion processing unit collects real-time data on vehicle location, remaining battery power, charging station status, and grid load through vehicle terminals, charging pile sensors, and road network and power grid SCADA systems. It then integrates and processes this data to build a basic data pool. The AI ​​model calculation unit, connected to the data fusion processing unit, constructs dynamic demand prediction models, multi-objective optimization scheduling models, and a power grid-charging station-truck collaborative control model. Based on these models, it calculates dynamic charging demand information, optimization information, and control paths for trucks. The scheduling decision unit generates optimal charging paths and scheduling strategies based on the model calculation results. It also receives execution results from the vehicle navigation system and charging pile control system after executing scheduling commands, and provides visualized monitoring of charging information for each truck and charging station data.

[0026] Specifically, such as Figure 2 As shown, the data fusion processing unit consists of a data preprocessing module and a multi-source data fusion module, forming a complete data processing link. The data preprocessing module interacts with the traffic management platform, power grid information system, and vehicle terminals through API interfaces to obtain the vehicle location, remaining battery power, connection status, output power, fault codes, substation load data, and real-time congestion index extracted from the traffic management platform API for each truck in the system. It performs outlier cleaning, spatiotemporal alignment, and data compression preprocessing on the data. Among these, the 3σ principle is used to detect and remove outliers from the vehicle battery state of charge (SOC) data, effectively eliminating abrupt changes in data caused by sensor failures or communication anomalies. For example, data with fluctuations greater than 20% within 10 seconds is removed. Spatiotemporal alignment unifies heterogeneous data to a spatiotemporal coordinate system composed of WGS-84 geographic coordinates and UTC timestamps. The Delta coding algorithm is used to compress vehicle trajectory data, reducing the storage of redundant trajectory points by approximately 60% while ensuring data accuracy. The multi-source data fusion module is used to extract features from the data, acquiring vehicle SOC information, current status information of charging piles, road network congestion index, electricity price time period, and grid load, and fusing them to generate a spatiotemporal feature matrix. Specifically, it uses the Kalman filter algorithm to fuse multi-sensor data to improve the accuracy of SOC estimation, extracts charging pile status (power, availability, fault codes) through one-hot encoding, and calculates dynamic congestion information by combining real-time vehicle speed and historical traffic flow using a historical similarity matching method, achieving accurate prediction of the dynamic traffic congestion index. For grid load, it uses wavelet decomposition to perform frequency domain analysis on the grid load data, extracting the fundamental frequency component to reflect the basic trend of grid load changes. Finally, it generates a feature matrix containing spatiotemporal dimensions, providing high-quality input data for subsequent model calculations.

[0027] The AI ​​model computation unit includes a demand forecasting module, an optimization scheduling module, a collaborative control and adjustment module, and a dynamic conflict resolution module. The demand forecasting module is used to construct a dynamic demand forecasting model based on a Long Short-Term Memory (LSTM) network. This dynamic demand forecasting model includes an input layer, a bidirectional LSTM layer, a spatiotemporal attention layer, a fully connected layer, and an output layer. The bidirectional LSTM layer captures the time-series features of charging demand, and the spatiotemporal attention mechanism enhances the model's ability to perceive regional differences in charging demand. Historical charging data, traffic flow, weather conditions, and electricity pricing policies are used as input features to the dynamic demand forecasting model to predict charging demand in the spatiotemporal domain. The output is the probability distribution of charging demand at each charging station within a preset future time period, thus improving the model's generalization ability.

[0028] In this embodiment, the input dimensions of the dynamic demand prediction model include historical charging parameters, temperature, and electricity price. The attention mechanism is used to calculate the regional charging demand weight Wi. The loss function used for model training is quantile loss to optimize long-tail distribution prediction. The training data is augmented by generating virtual charging scenarios under extreme weather conditions through a GAN network.

[0029] Specifically, the optimization scheduling module constructs a multi-objective optimization model with the goals of minimizing transportation costs, maximizing charging facility utilization, and minimizing grid load fluctuations. It then uses a particle swarm optimization algorithm to solve the multi-objective optimization model, generating an optimal charging route for each truck. Transportation costs include travel time and charging costs. The collaborative control and adjustment module implements a dynamic adjustment strategy based on collaborative control conditions, including real-time electricity price incentives, charging power adjustment, and truck route adjustments.

[0030] The real-time electricity price incentives in the coordinated control conditions include dynamically adjusting charging prices based on the real-time load status of the power grid using a price elasticity model. Prices are lowered during off-peak hours to attract trucks to charge, and raised during peak hours to suppress charging demand. For example, during peak grid load periods (load rate > 85%), the charging price is increased by 15%, and during off-peak periods (load rate < 30%), the price is reduced by 20%, guiding charging demand to shift to off-peak times. Charging power adjustment includes charging stations adjusting their output power in real-time according to the power grid load. For example, if multiple vehicles are connected to the same transformer simultaneously, the charging power is automatically and dynamically adjusted to 80kW to avoid power grid overload. Truck route adjustment includes replanning alternative charging routes for trucks based on the Dijkstra algorithm when the power grid load is too high or the charging queue time is too long, thus recommending alternative charging routes to avoid concentrated charging. For example, if the charging queue time is greater than 30 minutes, the truck is redirected to an adjacent station.

[0031] In this embodiment, the objective function of the multi-objective optimization model is: Where Ci is the transportation cost of the i-th truck, and Uj is the utilization rate of the j-th charging station. Let be the standard deviation of the grid load in time period t. , The weighting coefficients are used to balance the priorities of different objectives. The constraints of the objective function are: the remaining battery power of the vehicle is not lower than the safety threshold, the power of the charging station does not exceed the rated capacity, and the grid load does not exceed the peak limit.

[0032] The dynamic conflict resolution module establishes a charging pile competition monitoring mechanism to track conflict scenarios where multiple trucks approach the same charging pile in real time. If multiple vehicles approach the same charging pile, they are rescheduled according to their urgency. The urgency assessment includes SOC margin and task priority. The urgency assessment model is as follows: ,in, and These are the weighting coefficients. Remaining time for the task The deadline is determined by taking into account the vehicle's remaining state of charge (SOC) and the remaining time of the task. and deadline The system reallocates charging resources for vehicles experiencing conflicts. It also introduces a virtual charging queue mechanism to schedule temporary tasks for waiting vehicles, effectively preventing competition for charging resources and improving system efficiency.

[0033] In this embodiment, the scheduling decision unit consists of a charging path coordination module, an intelligent scheduling module, an execution feedback module, and a user interaction module, enabling precise execution of the scheduling strategy and user interaction management. The charging path coordination module receives the optimal charging path information for each truck and, based on the principle of dynamic programming and combined with real-time road condition information, optimizes the truck charging path in segments. By establishing a vehicle energy consumption model, considering factors such as road distance, gradient, and air conditioning usage, the energy consumption of the vehicle passing through each road segment is predicted. When the predicted remaining battery power is lower than a safety threshold (e.g., 15%) or cannot meet the demand to reach the next charging point, charging is arranged at the road segment node, with the charging amount based on the principle of meeting the subsequent travel needs and not exceeding 80% of the battery capacity, effectively shortening the charging time.

[0034] like Figure 3 As shown, the specific steps for segmented charging optimization include: Input the vehicle's initial SOC (State of Charge, current battery percentage) and the planned route (containing multiple route segments, each with a start point, end point, distance, and estimated energy consumption information); set the current battery level as the initial SOC and create an empty charging plan list; then iterate through each route segment: a. Calculate the energy consumption required to pass through the current segment, which can be estimated using the segment distance, vehicle energy consumption model (e.g., kWh / km), and environmental factors (e.g., gradient, air conditioning usage); b. Predict the remaining battery power after passing through the segment: soc_after_segment = current_soc - (required_energy / battery) c. Calculate the remaining battery capacity after passing through the current segment by multiplying it by 100% and converting the percentage of battery capacity into actual battery capacity (kWh); d. Check if the remaining battery capacity after passing through the current segment is lower than the safety threshold or lower than the battery capacity required to reach the next charging point (or destination) plus a redundancy, such as 10%. If so, charging should be arranged at the end of the current segment, i.e., the next node. When determining the charging amount, charge until the battery capacity is sufficient to meet the requirements of the subsequent segments to reach the next charging point (or destination) plus a safety redundancy, but not exceeding 80% of the battery capacity. Avoid fully charging to save time, as the battery charging curve slows down after 80%; e. Update the current battery capacity: if charging is done at this node, the current battery capacity becomes the charged battery capacity; if not, the current battery capacity is soc_after_segment; finally, continue traversing the next path segment until the entire path is completed, and output a charging plan list, which includes charging location and charging amount.

[0035] The intelligent scheduling module is connected to the charging path coordination module. The intelligent scheduling module converts path optimization results into executable instructions, which are used to send recommended charging stations and charging amounts to the vehicle navigation system based on the path optimization results, and to send adjustment parameters such as charging power and charging time to the charging pile control system, achieving precise allocation of charging resources. The execution feedback module establishes a two-way communication protocol to receive real-time execution status feedback information from the charging pile control system. Combined with charging data uploaded from the vehicle terminal, it achieves full-process monitoring of the truck charging process and charging station operation status through visual monitoring. The user interaction module adopts a role-based access control (RBAC) mechanism. Through authentication and permission allocation, it achieves hierarchical management of system administrators. It mainly obtains the identity information of administrators and compares it with personnel verification information in the basic database to determine whether they are legitimate personnel. Based on role information, it assigns corresponding operation permissions. It also supports users to perform query operations through a visual interface, such as viewing truck running trajectories, charging records, and real-time charging station status. Simultaneously, it allows users with appropriate permissions to edit and adjust system parameters (such as electricity price strategies and weighting coefficients), improving the flexibility and adaptability of system management.

[0036] This application also discloses a scheduling method based on the above system. This scheduling method relies on the collaborative operation of a data fusion processing unit, an AI model calculation unit, and a scheduling decision unit to achieve dynamic optimization of charging resources for new energy trucks. The specific steps are as follows: S1: Predict the charging demand density of each region in the next 2 hours. The data fusion processing unit first collects data such as vehicle location, remaining power, charging station status, and grid load in real time from multiple sources such as vehicle terminals, charging pile sensors, road network, and power grid SCADA systems, with a sampling period of 1 minute. After the data preprocessing module uses the 3σ principle to remove sudden changes in vehicle SOC data, it completes spatiotemporal alignment based on WGS-84 geographic coordinates and UTC timestamps, and uses Delta encoding to compress redundant trajectory points before transmitting the data to the AI ​​model calculation unit. The demand prediction module uses a dynamic demand prediction model built based on an LSTM neural network, taking multi-dimensional features such as historical charging data, traffic flow, weather conditions, and electricity pricing policies as input. The model mines time-series features through bidirectional LSTM layers, calculates regional charging demand weights Wi using a spatiotemporal attention mechanism to enhance the perception of demand differences in different regions, optimizes long-tail distribution prediction using the quantile loss function during training, and combines a GAN network to generate virtual charging scenarios under extreme weather conditions to enhance data diversity. Finally, it outputs the probability density distribution of charging demand in each region in a 100m×100m grid for the next 2 hours, providing a data foundation for subsequent scheduling decisions. S2: Generate an initial route-charging plan for each vehicle. Based on the charging demand prediction results of step S1, the optimization scheduling module constructs a multi-objective optimization model with the objectives of minimizing transportation costs, maximizing the utilization rate of charging facilities, and minimizing grid load fluctuations. The model is then solved, taking into account real-time road conditions and charging pile status. This generates an initial route-charging plan for each truck, including the optimal driving route, charging station selection, and charging quantity planning. S3: Real-time conflict monitoring and elastic rescheduling. The dynamic conflict resolution module monitors the relative position, remaining power, and task priority of each vehicle and charging pile in real time at 10-second intervals. When multiple trucks are detected approaching the same charging pile, the urgency of each vehicle is calculated based on the urgency assessment model. Vehicles with higher urgency are given priority in charging resources. At the same time, a virtual charging queue mechanism is introduced to plan temporary tasks for waiting vehicles, such as detouring to nearby empty charging stations or performing short-distance delivery tasks. In addition, the collaborative control and adjustment module continuously monitors the grid load status and charging station queuing status. When the grid load exceeds a preset threshold or the charging queuing time is greater than 30 minutes, path replanning is automatically triggered. Based on the Dijkstra algorithm, alternative charging paths are recommended for trucks to achieve spatial transfer and dynamic balance of charging load. The charging path coordination module of the scheduling decision unit simultaneously re-optimizes the single charging amount based on the new path information to ensure that vehicles can efficiently complete the charging process while meeting the transportation task requirements.

[0037] This invention constructs a spatiotemporal state network model that integrates driving paths and charging decisions, employs a distributed optimization algorithm to achieve multi-vehicle collaborative scheduling, and designs a dynamic conflict resolution mechanism to address competition for charging resources. This significantly improves fleet operating efficiency, reduces charging waiting time and total transportation costs, and is applicable to scenarios such as urban logistics and intercity freight.

[0038] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to the embodiments of this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A collaborative scheduling and optimization system for charging routes of new energy freight vehicles, characterized in that, include: Data fusion processing unit, AI model computing unit, and scheduling decision unit; The data fusion processing unit is used to collect vehicle location, remaining battery power, charging station status, congestion index and grid load data in real time through vehicle terminals, charging pile sensors, road network and power grid SCADA systems, and integrate and process the data to build a basic data pool. The AI ​​model calculation unit is connected to the data fusion processing unit. The AI ​​model calculation unit is used to construct a dynamic demand prediction model, a multi-objective optimization scheduling model, and a power grid-charging station-truck collaborative control model. Based on each model, it calculates dynamic demand information, optimization information, and control paths for truck charging. The scheduling decision unit is used to generate the optimal charging path and scheduling strategy based on the model calculation results, and to receive the execution results fed back after the vehicle navigation system and the charging pile control system execute the scheduling instructions, and to perform visual monitoring of the charging information of each truck and the charging station data.

2. The new energy freight vehicle charging route collaborative scheduling optimization system according to claim 1, characterized in that, The data fusion processing unit includes a data preprocessing module and a multi-source data fusion module; The data preprocessing module is used to obtain the vehicle location, remaining power of each truck in the system, the connection status, output power, fault codes, substation load data, and real-time congestion index extracted from the traffic management platform API, and to perform outlier cleaning, spatiotemporal alignment, and data compression preprocessing on the data. The multi-source data fusion module is used to extract features from the data, obtain vehicle SOC information, current status information of charging piles, road network congestion index, electricity price period and power grid load, and fuse them to generate a spatiotemporal feature matrix.

3. The new energy freight vehicle charging route collaborative scheduling optimization system according to claim 1, characterized in that, The AI ​​model computing unit includes a demand forecasting module, an optimization scheduling module, and a collaborative control and adjustment module; The demand forecasting module is used to construct a dynamic demand forecasting model based on an LSTM neural network. The dynamic demand forecasting model includes an input layer, a bidirectional LSTM layer, a spatiotemporal attention layer, a fully connected layer, and an output layer. Historical charging data, traffic flow, weather conditions, and electricity pricing policies are used as input features to the dynamic demand forecasting model to perform spatiotemporal forecasting of charging demand and output the probability distribution of charging demand for each charging station within a preset time period in the future. The optimization scheduling module is used to construct a multi-objective optimization model with the objectives of minimizing transportation costs, maximizing the utilization rate of charging facilities, and minimizing grid load fluctuations, and to solve the multi-objective optimization model to generate the optimal charging route for each truck. The transportation costs include travel time and charging costs. The collaborative control adjustment module is used to perform real-time electricity price incentives, charging power adjustment, and truck route adjustment based on collaborative control conditions.

4. The new energy freight vehicle charging route collaborative scheduling optimization system according to claim 3, characterized in that, The objective function of the multi-objective optimization model is: Where Ci is the transportation cost of the i-th truck, and Uj is the utilization rate of the j-th charging station. Let be the standard deviation of the grid load in time period t. , The weighting coefficients are used to define the constraints of the objective function: the vehicle's remaining battery power is not lower than the safety threshold, the charging station's power does not exceed the rated capacity, and the grid load does not exceed the peak limit.

5. The new energy truck charging route collaborative scheduling optimization system according to claim 4, characterized in that, The AI ​​model calculation unit also includes a dynamic conflict resolution module. This module monitors charging pile competition in real time. If multiple vehicles approach the same charging pile, they are rescheduled according to their urgency. The urgency assessment data includes SOC margin and task priority. The urgency assessment model is as follows: ,in, and These are the weighting coefficients. Remaining time for the task A deadline is set, and a virtual charging queue mechanism is introduced to plan temporary tasks for waiting vehicles.

6. The new energy freight vehicle charging route collaborative scheduling optimization system according to claim 1, characterized in that, The scheduling decision unit includes a charging path coordination module, an intelligent scheduling module, and an execution feedback module; The charging path coordination module receives the optimal charging path information for each truck and dynamically adjusts the single charging amount according to subsequent road conditions to optimize segmented charging. The intelligent scheduling module is connected to the charging path coordination module. The intelligent scheduling module is used to send charging station and charging amount recommendation information to the vehicle navigation system of the vehicle terminal based on the path optimization results, and to send charging adjustment parameters to the charging pile control system. The execution feedback module is used to receive the execution results fed back by the charging pile control system after executing the scheduling instructions, and to perform visual monitoring of the charging information of each truck and the charging station data.

7. The new energy freight vehicle charging route collaborative scheduling optimization system according to claim 1, characterized in that, The scheduling decision unit also includes a user interaction module, which is used to obtain the identity information of the management personnel, compare the obtained identity information with the personnel verification information in the basic database to determine whether they are legitimate personnel, and assign corresponding operation permissions according to role information. It is also used to obtain the user's query instructions and parameter editing instructions.

Images