AGV charging control system and control method
By introducing vehicle-pile mirror mapping and directed cross-position sequence into the AGV charging control system, stable charging under complex working conditions of multiple concurrent charging and frequent plugging and unplugging is achieved, solving the problems of single cell overvoltage and bus voltage surge, and ensuring the safety and efficiency of port roll-on/roll-off transportation.
Patent Information
- Application Number
- CN202511493930.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-20
- Publication Date
- 2026-02-27
- Estimated Expiration
- 2045-10-20
AI Technical Summary
In port roll-on/roll-off transportation, existing AGV charging control methods are difficult to effectively suppress overvoltage spikes in individual cells and bus voltage surges under conditions of low temperature, high SOC, individual cell consistency degradation, and frequent plugging and unplugging. Furthermore, they lack dynamic safety margins and adaptive constant voltage upper limits for individual cells, posing a risk of overcharging.
By establishing a vehicle-pile mirror mapping instance, and employing directed cross-position sequence and odd-even partition rotation, mutually exclusive bitmap and zero current, zero potential difference hard access, combined with the coordinated control of charging, unloading and regeneration power, we can achieve multi-vehicle and multi-pile non-intersection, predictable switching, and controlled plugging and unplugging, and superimpose segmented metering and consistent write-back.
It suppresses overvoltage spikes in individual cells and bus voltage surges, reduces the risk of interruption impacts and erroneous parallel connections, ensures the stability and safety of the charging process, supports stable execution under complex operating conditions, and provides consistency verification and auditing capabilities throughout the entire process.
Smart Images

Figure CN120942050B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of vehicle charging, in particular to an AGV charging control system and a control method. BACKGROUND
[0002] In port ro-ro transportation, the turnover intensity of commercial vehicles is high, the berthing window is short, the road and slope of the yard are complex, and the traditional tractor and manual driving have the problems of uneven scheduling, high energy consumption and prominent safety hazards. In order to improve the efficiency of loading and unloading and the level of zero emission, the existing research focuses on deploying all-terrain automatic guided vehicle (AT-AGV) with battery power source, environmental perception and autonomous navigation capability, through SLAM (Simultaneous Localization and Mapping) laser radar mapping, fusion positioning, path planning and precise docking to realize unmanned transfer of commercial vehicles (as shown in Figure 1 The existing charging control takes the battery pack voltage (Pack voltage) and state of charge (SOC) threshold of the battery management system (BMS) side as the basis for stopping charging, combines fixed hysteresis and constant current-constant voltage (CC-CV) two-stage strategy, and some systems add time-sharing power limit or start and stop according to task time window. However, under the port ro-ro transportation working conditions of low temperature, high SOC, single consistency degradation, and coexistence of frequent plug-in and regeneration backfilling, the AT-AGV is prone to single cell overvoltage peak in the CC to CV switching, plug-in instant and short-time traction stage after leaving the pole, the Pack voltage does not exceed the limit, but the local single cell is critical, which brings the risk of battery lithium precipitation and life attenuation; at the same time, multiple poles may cause direct current bus voltage overshoot and peak value exceeding limit; the existing control method based on fixed threshold and hysteresis strategy is difficult to feed forward and inhibit in time, lacks dynamic safety margin and adaptive constant voltage upper limit for single cell, and lacks cooperative control of plug-in soft retreat and regeneration power limiting, which is difficult to balance between safety, charging efficiency and power grid constraints, and has the risk of overcharging. SUMMARY
[0003] The technical problem to be solved by the present application is to provide an AGV charging control system and a control method, which realizes multiple vehicle and multiple pole non-intersection, predictable switching, controlled plug-in and suppression of single cell overvoltage and bus voltage overshoot through directed cross-site sequence and mirror mapping of vehicle and pole, cooperation of odd-even partition rotation, mutual exclusion bitmap and zero flow, zero potential difference hard access, and cooperation of charging, energy discharge, regeneration, limiting, admittance and soft retreat within the unified energy boundary, and superposition of segmented metering and consistency write-back.
[0004] To achieve the above purpose, the present application provides the following technical solutions:
[0005] An AGV charging control system, comprising a BMS, an AGV controller, a relay management module, a regenerative power interface, a charging power converter, a bus discharging module, a plug-in handshake device, a metering isolation unit and a charging management platform, the BMS provides cell and module states and voltage and current limits to the AGV controller, the AGV controller generates a directed cross-position sequence on a node array according to an occupancy mark and an access rule, determines a channel set, a parity sequence, a partition rotation order and a channel switching constraint, issues a motion instruction to the charging power converter and the bus discharging module, and synchronizes to the charging management platform, the charging management platform issues a charging pile mirror mapping table and a number, the plug-in handshake device binds a vehicle end number and a pile end number to form a mirror mapping instance after completing mechanical interlocking, identity authentication, insulation and metering isolation readiness, time synchronization, discharging standby and instruction arrival confirmation, the AGV controller issues a cross-position sequence to the charging power converter and the bus discharging module according to the mirror mapping instance, and issues an amplitude limiting and switching parameter to the regenerative power interface and the charging power converter, the relay management module controls the opening and closing of the pre-charging loop and the main loop according to the confirmation result and the time stamp, and the metering isolation unit steps to implement energizing, isolation and zero-flow verification according to the cross-position sequence, and writes the segmented energy metering, channel identification and time stamp back to the AGV controller and the charging management platform.
[0006] As a further scheme of the system, in the AGV controller, the occupancy mark is composed of the contact state of the relay management module, the channel occupancy identification of the metering isolation unit, the bus potential difference and current zero-flow mark, the mechanical interlocking and authentication mark of the plug-in handshake device; the access rule is composed of a preset fault isolation table, an insulation unready table, an identity inconsistency table, a temperature out-of-limit table, an SOC upper limit table, a voltage upper limit table and a maintenance locking table; the node array is composed of a vehicle end node, a pile end node and a bus segmented node, wherein the vehicle end node includes a pre-charging loop node, a main loop node, a regenerative loop node and a Pack terminal node, the pile end node includes a power converter output node, a metering node, an isolation node and a discharging node; the channel set is composed of a pre-charging channel, a main charging channel, a discharging channel and a regenerative amplitude limiting channel and its allowed connection edge; the parity sequence is a parity sequence in which the stepping index and the landing point index are alternately arranged according to the time index; the partition rotation order is a partition rotation order of a front area, a buffer area and a standing column area divided according to the yard area and the bus segment; the channel switching constraint includes a zero-flow criterion, a zero-potential difference criterion, a pre-charging prior to main loop closure, a soft retreat prior to disconnection, a relay debounce time window and a cross-channel non-parallel rule set.
[0007] As a further scheme of the system of the application, the BMS indexes and time marks the sampling data of the battery cell and the module, and generates an occupancy mark word containing the relay contact mirror, the metering isolation mirror, the zero current criterion and the zero bus potential difference criterion; generates an exclusion rule word according to the fault isolation item, the insulation unready item, the identity inconsistency item, the maintenance lock item, the temperature window item and the SOC and voltage upper limit item; calculates the limit value parameter set of the battery cell voltage limit, the module voltage limit, the Pack constant voltage target, the current upper limit and the regenerative power upper limit according to the temperature, the SOC, the aging identifier, the charge-discharge rate and the charge stage identifier, and outputs the battery cell index and the sorting vector of the battery cell whose voltage is lower than the preset voltage threshold, the temperature is higher than the preset temperature threshold, the SOC is lower than 20% or the SoH is lower than 70%, the preset voltage threshold and the preset temperature threshold are determined according to the battery cell attribute parameter, the battery cell index is the index position of the battery cell in the battery pack, and the sorting vector is calculated by weighting the comprehensive score of the voltage, the temperature, the SOC and the SoH of the battery cell, and the comprehensive score is sorted from small to large to generate the battery cell sorting vector; estimates the first and second order change rates of the battery cell voltage, and forms the slope and curvature parameters for the prediction current limiting; packs the occupancy mark word, the exclusion rule word, the limit value parameter set, the battery cell index meeting the preset threshold requirement and the time mark into a message and publishes it to the AGV controller with the vehicle pile mirror mapping instance number.
[0008] As a further scheme of the system of the application, the charging power converter receives the charging power conversion action instruction frame issued by the AGV controller according to the directed cross-position sequence, the charging power conversion action instruction frame includes step number, channel ID, constraint ID, timestamp, CC target current, CV target voltage, charging current change rate and charging voltage change rate limit, limit current power-on in pre-charge channel, drive CC trajectory according to the table in main charge channel and implement constant voltage switching according to the charging current change rate and the charging voltage change rate limit when the switching condition is reached, start and stop the power channel according to the odd-even arrangement and partition rotation time slot, reduce the output current to zero according to the zero flow time sequence when the soft retreat trigger or the plug-in node zero flow criterion is received, and keep the bus and output end potential consistent, check the channel switching constraint according to the constraint ID, and perform rejection and state latching on inconsistent instructions, and return the voltage, current, channel state, energy metering and timestamp to the AGV controller and the charging management platform according to the segment number.
[0009] As a further embodiment of the system of the present invention, the bus unloading module receives unloading instruction frames issued by the AGV controller in a directed cross-position sequence. The unloading instruction frame includes step number, bus segment number, constraint ID, timestamp, target bus potential, equivalent admittance curve, power limit, charging power change rate and charging voltage change rate limit, activation / deactivation threshold, and dwell time slot. In the designated bus segment, the electronic load channel or resistor channel is activated according to the equivalent admittance curve, and the unloading trajectory is formed according to the charging power change rate and charging voltage change rate limit. The channel activation / deactivation is controlled according to the odd-even arrangement and partitioned rotation time slot. The unloading pre-setup and exit sequence is executed according to the plug-in node zero current criterion and metering isolation ready state. The non-crossing and channel switching constraints are verified according to the constraint ID, and the inconsistent instructions are rejected and the status is latched. The bus potential, unloading current, energy metering, channel identifier and timestamp are reported to the AGV controller and charging management platform according to the segment number.
[0010] As a further embodiment of the system of the present invention, the regenerative power limiting interface receives regenerative control frames issued by the AGV controller based on a mirror mapping instance. The regenerative control frame includes a step number, channel identifier, constraint identifier, timestamp, and a regenerative power upper limit table, DC bus voltage upper limit table, braking torque limiting table, power and current ramp-up slope, short-time suppression threshold, and dwell time slot indexed by minimum safety margin, temperature, state of charge, and health status. Based on the channel identifier mapping, it generates motor braking torque limiting and DC side current limiting, and outputs them through rate limiting and feedforward filtering shaping. It enables or freezes the limiting in the corresponding time slot according to odd / even alignment and partitioned rotation. Upon receiving a zero-current criterion for plug-in / plug-out nodes or a status change of the relay management module, it enters the soft limiting stage and performs cancellation or restoration according to the timing table. It completes the verification of non-crossing and channel switching constraints according to the constraint identifier, and performs rejection and status latching for inconsistent instructions. It reports the regenerative power, DC current, motor braking torque, channel identifier, and timestamp to the AGV controller and charging management platform according to segment numbers.
[0011] As a further embodiment of the system of the present invention, the metering isolation unit establishes a binding relationship between segmented metering channels and channel identifiers using the cross-sequence step number as the main index. It completes the zero-current criterion triggering by combining dual-channel current sampling and zero-sequence detection. It performs three-state switching of power-on, disconnection and bypass on the vehicle-end and pile-end channels according to the programmable isolation matrix. It performs numbering and cyclic redundancy verification on the metering frame and isolation frame based on the mirror mapping instance and writes the timestamp. It sets the hierarchical isolation priority and interlocking conditions of the bus section and branch section. It coordinates the opening and closing sequence of the relay management module through edge-triggered zero-current interlocking and potential difference interlocking. It completes the segmented energy metering accumulation and metering segment number update in steps, and writes back the metering data, isolation status word, channel identifier and timestamp to the AGV controller and charging management platform in the form of a message.
[0012] The application discloses an AGV charging control method and system.
[0013] Step 1, mapping and reference are established: mirror mapping instances are generated by acquiring the car end and stake end numbers of the AGV, time reference, serial number and occupation code initialization are completed, monomer cell voltage, module end voltage and Pack total voltage, DC bus current, charging output current and regeneration backfill current, cell temperature, module temperature and Pack shell temperature, pre-charge contact and main contact state and plug-in interlock state are collected, bus potential difference and zero flow flag are acquired, occupation mark word and forbidden entry rule word are formed;
[0014] Step 2, sequence arrangement and rotation: a directed cross-position sequence is generated on the discrete node set according to the occupation mark and forbidden entry rule, the allowed connection edge, step sequence and parity sequence are determined, and the candidate steps are loaded into the partition queue according to the region and bus segmentation, and the current execution step is selected through the rotation strategy;
[0015] Step 3, constraint checking and channel access: non-intersection, zero flow, zero potential difference, de-bouncing time window and channel switching sequence are matched and stored, the target channel is enabled or isolated and zero flow is checked, and the channel identifier and metering segment are recorded;
[0016] Step 4, charging power parameterization and energy discharge coordination: charging constant current target, charging constant voltage target, charging power change rate boundary, charging amplitude limiting parameter and residence time slot are generated for the current step, and the parameters are downloaded, when the charging energy is withdrawn, the equivalent charging admittance curve, charging power and voltage change rate boundary, charging on-off threshold and residence time slot are generated synchronously;
[0017] Step 5, regeneration amplitude limiting and time slot control: the regeneration power, DC side current and motor torque limiting are calculated according to the minimum safety margin of the monomer cell during charging, temperature and state of charge index, the limiting is enabled or frozen according to the parity and partition time slot, and when the plug-in zero flow or relay management state changes, the limiting is removed or restored in the soft limiting stage;
[0018] Step 6, execution confirmation, promotion and constraint archiving: zero flow, zero potential difference, parameter readiness and consistency are confirmed and time stamp and event sequence are written, when the conditions are met, the next step is promoted, when the conditions are not met, the path and parameters are rearranged, after reaching the termination node, soft return, channel withdrawal and disconnection are performed in sequence, and metering is written back and record archiving is completed.
[0019] As a further solution of the method of the present application, in step 2, the sequence arrangement includes: representing the discrete node set as a directed graph G(V, E), with V consisting of node types, busbar segment numbers and area identifications, to allow connection edges and their channel types to constitute E; performing filtering on E with occupancy markers and forbidden rules to obtain E'; attaching step sequence indexes, parity markers, channel IDs and time stamp windows to each edge; using adjacency tables and bitmap to latch the occupancy states of nodes and edges; generating a directed cross-bit sequence on the subgraph (V, E') by restricted topology traversal with the starting node set S stably ordered in lexicographic order and time stamp as the root, and performing precondition checking, edge label consistency checking and time window consistency checking on each step during the traversal, and writing the sequence in a two-stage reservation and commit process.
[0020] As a further solution of the method of the present application, in step 2, the round-robin selection includes: writing candidate steps into partition queues {Q K} according to areas and busbar segments, each queue recording a queue token, a remaining weight and a last service time; selecting {Q K} using a gap round-robin algorithm, limiting the parity markers of available steps according to a parity mapping table during selection, and performing conflict detection using mutual exclusion bitmaps of channel IDs and node IDs; performing a two-phase commit of reservation-confirmation on the candidate steps that pass the detection, setting lock bit flags of nodes and edges in the reservation phase, and completing the commit according to the time stamp window and constraint ID in the confirmation phase; when there is concurrent competition, making a decision according to the ternary tuple {priority, time stamp, sequence number}; updating the weights of candidate steps that are not selected according to aging rules and preserving the order of entry for the next round of selection.
[0021] As a further solution of the method of the present application, the present application can solve the problems of lack of unified timing, energy boundary, dynamic safety margin and concurrent scheduling, difficulty in suppressing switching transients and channel conflicts, and easy overvoltage of single cells and busbar surge in the port charging scenario with Pack threshold and CC-CV control.
[0022] The technical effect of the present application: by establishing the mirror mapping instance of the vehicle end and the pile end and the unified time reference, the reported occupation mark and the no-entry rule drive the generation and execution of the directed cross-position sequence, the admission, parameter download and relay opening and closing time sequence arrangement of each channel are implemented under the odd-even arrangement and partition rotation mechanism, and the gradual stable confirmation and violation blocking are completed with zero flow and zero potential difference criterion, de-bouncing time window and channel switching constraint, so that the charging power converter realizes the smooth switching from constant current to constant voltage according to the limit value, the bus energy withdrawal module realizes the energy withdrawal shaping according to the equivalent admittance curve, and the regeneration side implements the power, current and torque limiting according to the minimum safety margin; the metering isolation unit writes back the segmented energy, channel identifier and time stamp, and the event sequence is used for path rearrangement and consistency checking, so as to suppress the overvoltage transient of single cell and the bus voltage surge under the conditions of multi-pile concurrent and frequent plug-in, reduce the opening impact and misparallel risk, and guarantee the stability of the charging process and the traction task. At the same time, through the association of segmented metering and mirror number, a traceable audit link is formed for checking and review, and the parameterized generation of cell index and change rate boundary meeting the preset threshold requirement is matched, which reduces the current limiting false trigger and is compatible with low temperature, high SOC and aging difference working conditions, supports scheduling window charging and online update of strategy, meets the bus power and power quality constraints, and reduces the risk of AGV overcharging. BRIEF DESCRIPTION OF DRAWINGS
[0023] Figure 1 The present application AGV existing equipment diagram.
[0024] Figure 2 The present application is an AGV charging control system interface diagram.
[0025] Figure 3 The flow chart of the control method of the present application.
[0026] Figure 4 The technical route diagram for solving the problem that when AGV is downhill or deceleration return regeneration and charging coexist, the regeneration backfill may push up the bus voltage and reach the CV upper limit. Traditional control is independent processing, lacking of collaborative and time slot management technology for charging and energy withdrawal.
[0027] Figure 5 The present application is a directed graph example diagram. DETAILED DESCRIPTION
[0028] The technical solutions in the embodiments of the present application will be described clearly and completely in the embodiments of the present application combined with the drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0029] Example 1
[0030] Figure 1 For example, a diagram of existing AGV equipment. Figure 2 As shown, the present invention proposes an AGV charging control system, including a BMS, an AGV controller, a relay management module, a regenerative power interface, a charging power converter, a bus de-energizing module, a plug-in / plug-out handshake, a metering isolation unit, and a charging management platform. The BMS provides the AGV controller with the status of the battery cells and modules and voltage and current limits. The AGV controller generates a directed spanning sequence on the node array based on occupancy markers and prohibition rules, and determines the channel set, odd-even arrangement, partition rotation order, and channel switching constraints. It then sends action commands to the charging power converter and the bus de-energizing module, and synchronizes them to the charging management platform. The charging management platform issues a vehicle-charging station mirror mapping table and... After the numbering and plug-in handshake completes mechanical interlocking, identity authentication, insulation and metering isolation readiness, time synchronization, energy unloading standby, and command arrival confirmation, it binds the vehicle-end number and the pile-end number to form a mirror mapping instance. The AGV controller issues the cross-position sequence to the charging power converter and the bus unloading module according to the mirror mapping instance, and issues the limiting and switching parameters to the regenerative power interface and the charging power converter. The relay management module controls the opening and closing of the pre-charging circuit and the main circuit according to the confirmation result and timestamp. The metering isolation unit performs empowerment, isolation and zero-current verification step by step according to the cross-position sequence, and writes back the segmented energy metering, channel identification and timestamp to the AGV controller and the charging management platform.
[0031] Compared to existing charging control methods that rely on Pack voltage, SOC threshold, fixed hysteresis, and simple CC-CV switching, this invention employs a directed cross-position sequence to discretize handshake, pre-charging, main charging, CV, soft deactivation, and energy dissipation into a step-by-step process with preconditions and time windows. It uses a vehicle-pile mirror mapping instance to achieve one-to-one binding of identity with commands, metering, and events; it uses odd-even allocation, partitioned rotation, and mutually exclusive bitmaps for concurrent step selection; it uses zero-current and zero-potential-difference criteria, along with debouncing timing, as electrical access and convergence thresholds; it uses energy dissipation based on equivalent admittance activation / deactivation and regeneration limiting based on minimum safety margin indexing, time slot enabling, and freezing to form a three-domain coordination; and it uses segmented metering bound by cross-position step numbers + CRC write-back to form a traceable link. This enables trajectory constraints on energy surges during CC→CV switching, plugging / unplugging, and recharge processes, suppressing individual overvoltage spikes and bus surges, avoiding channel crossings and erroneous parallel connections when multiple vehicles and multiple piles operate concurrently, reducing interruption impacts and ensuring consistent switching sequence. It supports stable execution under low temperature, high SOC, and frequent plugging / unplugging conditions, and provides consistency verification and auditing capabilities throughout the entire process.
[0032] It should be noted that in the AGV controller, the occupation flag is an occupation flag composed of the contact state of the relay management module, the channel occupation identification of the metering isolation unit, the bus potential difference and current zero flow flag, the mechanical interlocking and authentication flag of the plug-in handshaker; the forbidden entry rule is a forbidden entry rule composed of a preset fault isolation table, an insulation not ready table, an identity inconsistency table, a temperature out-of-limit table, a SOC upper limit table, a voltage upper limit table, and a maintenance locking table; the node array is a node array composed of a vehicle end node, a pile end node, and a bus section node, wherein the vehicle end node includes a pre-charging loop node, a main loop node, a regeneration loop node, and a Pack terminal node, the pile end node includes a power converter output node, a metering node, an isolation node, and a discharging node; the channel set is a channel set composed of a pre-charging channel, a main charging channel, a discharging channel, and a regeneration limiting channel and its allowed connection edge; the parity array is a parity array that alternately arranges step indexes and landing indexes according to time indexes; the partition rotation order is a partition rotation order of a front area, a buffer area, and a standing column area divided according to a yard area and a bus section; the channel switching constraint includes a zero flow criterion, a zero potential difference criterion, pre-charging before main loop closure, soft retreat before disconnection, a relay debounce time window, and a cross-channel non-parallel rule set. By discretizing the charging process into node-channel-step cross-position arrangement, and uniformly constraining the occupation flag, the forbidden entry rule, the parity array, and the partition rotation, in combination with the zero flow and zero potential difference criteria and the debounce time window, the pre-charging before the main loop closing and the soft retreat before the disconnection switching rule, the channel is not crossed and the identity is not confused under the multi-car and multi-pile concurrency, the single cell overvoltage and bus surge caused by switching and plug-in transient are suppressed, the consistency of the contactor action sequence and parameters under the installation is guaranteed, and the reviewable time sequence anchor point is provided for the metering write-back and event tracking.
[0033] It needs to be explained that the BMS indexes and time marks the sampling data of the battery cell and the module, and generates an occupancy mark word containing the relay contact mirror, the metering isolation mirror, the zero current criterion and the zero bus potential difference criterion; According to the fault isolation item, the insulation is not ready item, the identity inconsistency item, the maintenance lock item, the temperature window item and the SOC and voltage upper limit item, generate the forbidden rule word; According to the temperature, SOC, aging identification, charge-discharge rate and charge stage identification, calculate the limit value parameter set of the battery cell voltage limit, module voltage limit, Pack constant voltage target, current upper limit and regenerative power upper limit, and output the battery cell index and sorting vector whose voltage is lower than the preset voltage threshold, temperature is higher than the preset temperature threshold, SOC is lower than 20% or SoH is lower than 70%, the preset voltage threshold and the preset temperature threshold are determined according to the battery cell attribute parameters (the preset voltage threshold and the preset temperature threshold are determined according to the battery cell type, the rated voltage, the rated temperature range and the design specification of the battery, the voltage threshold is usually set to 90% to 95% of the maximum rated voltage of the battery cell, to ensure the safe working range of the battery, when the battery is just put into use; The temperature threshold is determined according to the thermal stability and maximum working temperature of the battery cell material, usually set to the upper limit of the working temperature of the battery cell minus 10 ℃ to 15 ℃, to prevent the battery from being damaged or reducing performance due to overheating, when the battery is just put into use, the voltage threshold is set to 95% of the maximum rated voltage, to ensure that the battery can be charged with high efficiency in the initial stage of normal use, and the temperature threshold is set to the maximum working temperature minus 10 ℃, to ensure that the battery will not be damaged due to overheating in the initial stage of starting, with the increase of use time and the aging of the battery, the voltage threshold is gradually adjusted to 90%, and the temperature threshold is gradually reduced to the maximum working temperature minus 15 ℃, to protect the battery from long-term load and temperature fluctuation), the battery cell index is the index position of the battery cell in the battery pack, and the sorting vector is calculated according to the weighted score of the voltage, temperature, SOC and SoH of the battery cell, and the comprehensive score is sorted from small to large to generate the battery cell sorting vector; The first and second order change rate of the battery cell voltage is estimated, and the slope and curvature parameters for predicting current limiting are formed; The occupancy mark word, forbidden rule word, limit value parameter set, battery cell index meeting the preset threshold requirement and time mark are packaged into a message and published to the AGV controller with the vehicle pile mirror mapping instance number. By completing the frame indexing and time marking of the sampling data at the source end, the states of the relay contact, metering isolation, zero flow and zero bus potential difference are compressed into the occupancy mark word, and the forbidden rule word is generated from the fault, insulation, identity, maintenance, temperature and SOC, voltage item, combined with temperature, SOC, aging and rate, stage to form the layered limit value of battery cell, module and Pack, and the battery cell index meeting the preset threshold requirement, and then the first and second order estimation of voltage is given to give the prediction current limiting slope and curvature, together with the mirror mapping number, as a message with time anchor point published to the control end, thereby providing a unified, feedforward and traceable parameter benchmark and consistency basis for the access, limiting and switching of cross-position sequence.
[0034] It should be noted that the charging power converter receives the charging power transformation action instruction frame issued by the AGV controller according to the directed cross sequence, the charging power transformation action instruction frame includes step sequence number, channel ID, constraint ID, timestamp, CC target current, CV target voltage, charging current rate and charging voltage rate limit, soft return trigger flag, limit current power-on in pre-charge channel, drive CC trajectory according to table in main charge channel and implement constant voltage switching according to charging current rate and charging voltage rate limit when switching condition is reached, start and stop power channel according to odd-even arrangement and partition rotation time slot, reduce output current to zero according to zero flow timing when soft return trigger or plug-in node zero flow criterion is received, and keep bus and output end potential consistent, check channel switching constraints according to constraint ID, and perform rejection and state latching on inconsistent instructions, and report voltage, current, channel state, energy metering and timestamp to AGV controller and charging management platform according to segment number.
[0035] It should be noted that the bus energy discharge module receives the energy discharge instruction frame issued by the AGV controller according to the directed cross sequence, the energy discharge instruction frame includes step sequence number, bus segment number, constraint ID, timestamp, target bus potential, equivalent admittance curve, power upper limit, charging power rate and charging voltage rate limit, switching threshold, and resident time slot, and the electronic load channel or resistance channel is put into the specified bus segment according to the equivalent admittance curve, and the energy discharge trajectory is formed according to the charging power rate and the charging voltage rate limit, the channel is controlled to be put in and taken out according to the odd-even arrangement and the partition rotation time slot, the pre-positioning and exit timing of the energy discharge is executed according to the plug-in node zero flow criterion and the metering isolation ready state, the non-crossing and channel switching constraints are checked according to the constraint ID, and the rejection and state latching are performed on the inconsistent instructions, and the bus potential, energy discharge current, energy metering, channel identification and timestamp are reported to the AGV controller and the charging management platform according to the segment number. By controlling the put-in and take-out according to the equivalent admittance curve in the target bus segment and being bound by the power and voltage rate limits, the channel start and stop is arranged according to the odd-even and partition time slots, the pre-positioning and exit timing control of the plug-in zero flow and metering isolation readiness is superimposed, and the online checking and violation latching of the non-crossing and channel switching constraints are implemented, a predictable energy discharge trajectory and unified energy boundary are constructed, the bus surge and backfill superposition caused by switching and plug-in are suppressed, concurrent channel conflicts are avoided, and the potential, current and energy data reported according to the segment number form a traceable consistency baseline, so as to maintain stable advancement of the cross sequence in cooperation with the charging power download.
[0036] It should be noted that the regenerative power limit interface receives the regenerative control frame issued by the AGV controller based on the mirror mapping instance, the regenerative control frame includes step sequence number, channel identifier, constraint identifier, time stamp, regenerative power upper limit table according to minimum safety margin, temperature, state of charge and health state index, direct current bus voltage upper limit, brake torque limiting table, power and current climbing slope, short time suppression threshold and residence time slot; According to the channel identifier mapping, the motor brake torque limiting and direct current limiting are generated, and the output is shaped through rate limiting and feedforward filtering; According to the parity alignment and partition rotation, the limiting is enabled or frozen in the corresponding time slot; When receiving the plug-in node zero flow criterion or the state change of the relay management module, enter the soft limiting stage, and execute the cancellation or recovery according to the time sequence table; According to the constraint identifier, the non-crossing and channel switching constraints are checked, and the inconsistent instructions are executed to refuse and state latch; The regenerative power, direct current, motor brake torque, channel identifier and time stamp are reported to the AGV controller and charging management platform according to the segmented number. Through the regenerative control frame bound by the mirror instance, the table is generated according to the minimum safety margin, temperature, SOC, SOH in the corresponding channel time slot, and the motor brake and direct current limiting are shaped through rate limiting and feedforward filtering, combined with the soft limiting time sequence triggered by the plug-in zero flow and the relay state, and the online checking and exception locking of non-crossing and channel switching constraints, the predictable and controllable regenerative power boundary is formed, the bus voltage surge caused by backflow is inhibited, and the concurrent conflict is inhibited, and the regenerative power, direct current and torque are reported by segmented number to realize the consistency and traceability of the whole process, so as to ensure that the cross sequence is stable according to the rules.
[0037] It needs to be explained that the metering isolation unit is indexed by the cross-position sequence step number, the segmented metering channel and the channel identification are bound, the zero flow criterion trigger is completed by combining double-channel current sampling and zero sequence detection, the power-on, disconnection and bypass three-state switching of the vehicle end and the pile end channel is performed according to the programmable isolation matrix, the metering frame and the isolation frame are numbered and cyclic redundancy checked based on the mirror mapping instance, and the time stamp is written, the hierarchical isolation priority and the interlocking condition of the bus section and the branch section are set, the opening and closing sequence of the relay management module is coordinated through the edge triggered zero flow interlocking and the potential difference interlocking, the segmented energy metering accumulation and the metering section number update are completed in steps, and the metering data, the isolation state word, the channel identification and the time stamp are written back to the AGV controller and the charging management platform in the form of a message. By establishing the metering section-channel identification binding with the cross-position step number and triggering the zero flow criterion by double-channel sampling+zero sequence detection, the vehicle end and the pile end channel are implemented power-on, disconnection and bypass three-state sequence by the programmable isolation matrix, the frame level consistency and traceability are realized by cooperating mirror numbering and CRC and time stamp, the hierarchical isolation priority and interlocking condition of the bus section and the branch section are adopted, and the zero flow and potential difference interlocking are used to coordinate the opening and closing sequence of the relay, so that the energy increment and the isolation state of each step are written back by section, forming an auditable metering baseline and providing a reliable data anchor point for the access review and path rearrangement of the cross-position sequence.
[0038] Embodiment 2
[0039] The difference between the embodiment 2 and the embodiment 1 of the application is that the embodiment 2 introduces an AGV charging control method.
[0040] As shown in Figure 3 The AGV charging control method provided by the application comprises the following steps:
[0041] Step 1, mapping and reference establishment: generating a mirror mapping instance by obtaining the vehicle end and pile end number of the AGV, initializing the time reference, sequence number and occupation code, collecting the single cell voltage, module end voltage and Pack total voltage, direct current bus current, charging output current and regenerative backfill current, cell temperature, module temperature and Pack shell temperature, pre-charging contact and main contact state and plug-in interlocking state, and obtaining the bus potential difference and zero flow flag, forming the occupation mark word and the forbidden rule word;
[0042] Step 2, sequence arrangement and rotation: generating a directed cross-position sequence on a discrete node set according to the occupation mark and the forbidden rule, determining the allowed connection edge, step sequence and parity arrangement, and loading the candidate steps into the partition queue according to the region and bus section, and selecting the current execution step through the rotation strategy;
[0043] Step 3, constraint check and channel access: rule matching and state latching are performed on non-crossing, zero-flow, zero-potential difference, de-bouncing time window and channel switching sequence, and enablement or isolation is implemented on the target channel, as well as zero-flow check, recording channel identification and metering section;
[0044] Step 4, charging power parameterization and discharging coordination: charging constant current target, charging constant voltage target, charging power change rate boundary, charging amplitude limiting parameter and residence time slot are generated for the current step, and parameters are downloaded, when involving charging energy withdrawal, equivalent charging admittance curve, charging power and voltage change rate boundary, charging on-off threshold and residence time slot are synchronously generated;
[0045] Step 5, regeneration amplitude limiting and time slot control: regeneration power, direct current side current and motor torque amplitude are calculated according to the minimum safety margin of single cell, temperature and state of charge index during charging, and the amplitude is enabled or frozen according to odd-even and partition time slot, and when the zero-flow or relay management state changes, the amplitude is removed or restored in the soft limiting stage;
[0046] Step 6, execution confirmation, promotion and constraint archiving: zero-flow, zero-potential difference, parameter readiness and consistency are confirmed and time stamp and event sequence are written, when the conditions are met, it is promoted to the next step, when it is not met, it is frozen and the path and parameters are rearranged, after reaching the termination node, soft return, channel withdrawal and disconnection are performed in sequence, and metering is written back and record archiving is completed.
[0047] The existing charging control stops at the Pack voltage and SOC threshold, cooperates with fixed hysteresis and two-stage CC-CV, and roughly starts and stops according to the task window, lacks unified arrangement for single cell margin, concurrent conflict and plug-in, and backfill transient, and the pile and vehicle association is mainly based on physical port or simple session, and the metering is mainly total amount without segmented traceability; this method uses directed cross-position sequence discrete handshake, pre-charging, main charging, CV, soft return and discharging, and realizes one-to-one binding of vehicle and pile and instruction and metering routing through mirror mapping instance, uses occupation mark + exclusion rule as pre-screening, supplemented by odd-even arrangement + partition rotation + mutual exclusion bitmap scheduling concurrent steps, uses zero-flow and zero-voltage change + de-bouncing as channel access and convergence threshold, downloads constant current and constant voltage targets and change rate boundary on the power side, and cooperates with segmented metering + isolation matrix + time stamp, CRC write back to form traceable link, so as to constrain the energy transition of CC→CV switching, plug-in and backfill to a predictable trajectory, suppress the overvoltage peak of single cell and the bus voltage surge, avoid the channel crossing and mis-parallel connection of multiple vehicles and multiple piles, ensure the consistency of contactor timing and parameters, and maintain stable promotion and compliance audit under low temperature, high SOC and frequent plug-in working conditions.
[0048] It should be noted that in step 2, the sequence arrangement includes: representing the discrete node set as a directed graph G(V, E), with V consisting of node types, bus segment numbers and area identifiers to allow connection edges and their channel types to consist of E; filtering E with occupancy markers and access rules to obtain E'; attaching step sequence indexes, parity markers, channel IDs and timestamp windows to each edge; using adjacency tables and bitmaps to latch the occupancy status of nodes and edges; generating a directed cross-bit sequence on the subgraph (V, E') using a restricted topology traversal with a stable order of the starting node set S in lexicographic order and timestamp, and performing precondition checks, edge label consistency checks and time window consistency checks for each step during traversal, and writing the sequence in a two-stage reservation and commit process.
[0049] By formalizing the node and channel relationship as a directed graph and filtering the feasible edge set in real time with occupancy markers and access rules, and attaching step sequence, parity, channel ID and time window to the edges and latching the occupancy status with adjacency tables and bitmaps, and combining the restricted topology traversal with stable ordering of the starting set and the two-stage reservation-commit writing mechanism, the generation of the cross-bit sequence has a determined sequence and reproducibility, which can avoid path intersection and resource preemption in a concurrent scenario, ensure precondition, label and time window consistency, reduce the risk of deadlock and starvation, and facilitate online tracking, auditing and rearrangement and expansion.
[0050] As shown in Figure 5 AGV charging control directed graph G with 20 nodes, where each node represents a charging control step (identified as "node" in the figure), and the edges represent the flow relationship between steps. Each step in the graph is equipped with specific indicators describing its relevant parameters, such as "channel" and "timestamp". The edge between "node 1" and "node 2" indicates an "odd time slot", meaning that in a specific time slot, the control system will select a specific channel for charging or discharging operations. The "channel ID" and "timestamp" attached to each edge indicate the execution time range and corresponding control channel for that step. For example, the edge from "node 1" to "node 2" indicates that channel 1 will be activated between 0 and 1000 milliseconds, while the edge from "node 2" to "node 3" uses channel 2 to control between 1001 and 2000 milliseconds. In addition, the indicators between steps also include "parity marker" and "channel identifier", such as the edge between "node 1" and "node 3", which indicates the use of "channel 3" in "odd time slots". The "timestamp" defines the specific time range for the execution of each step, ensuring that the system performs each task according to strict time management and step order. The "start" and "end" nodes in the graph represent the start and end of the AGV charging control process, respectively. The charging control system determines the specific tasks to be performed in each time period based on the reasonable combination of steps and time slots. Figure 5Through dynamic time slot management and accurate channel selection, the AGV charging, regenerative feedback, and energy unloading processes are ensured to run in a safe and efficient framework.
[0051] It should be noted that in step 2, the round-robin selection includes: writing the candidate steps into the partition queue {Q K} according to the area and bus section segmentation; each queue records the queue token, the remaining weight, and the last service time; the gap round-robin algorithm is used to select {Q K}, and during selection, the parity label of the available step is limited according to the parity mapping table, and the conflict detection is performed using the mutual exclusion bitmap of the channel ID and the node ID; the two-phase commit of reservation-confirmation is performed on the candidate steps that pass the detection, the lock bit flag of the node and the edge is set in the reservation phase, and the commit is completed according to the timestamp window and the constraint ID in the confirmation phase; when there is concurrent competition, the decision is made according to the three-tuple of {priority, timestamp, sequence number}; the weight of the unselected candidate step is updated according to the aging rule, and the order of entering the queue is retained for the next round of selection.
[0052] By constructing the partition queue according to the area and bus section segmentation and selecting the step under the parity constraint by gap round-robin, combined with the concurrent conflict detection and the two-phase commit of reservation-confirmation using the mutual exclusion bitmap of the channel and the node, as well as the decision of {priority, timestamp, sequence number}, the aging weighting of the unselected step, and the retention of the order of entering the queue, the execution order is fair, non-crossing, and predictable under the condition of multiple vehicles and multiple piles, avoiding resource preemption and starvation, reducing switching jitter and shortening waiting time, improving channel utilization and time slot hit rate, and maintaining the determinism, traceability, and auditability of sequence advancement under the consistency check of the time window and the constraint ID.
[0053] The technical solutions of the present application described in combination with Embodiments 1 and 2 are described below in combination with specific technical problems to describe the problem solving process.
[0054] (1) At room temperature, single-vehicle opportunity charging may still cause transient disturbance at the junction of pre-charging-main loop switching and CC→CV. Traditional Pack threshold and fixed hysteresis control lack hard access to zero flow, zero potential difference, and action sequence, leading to engineering hazards such as inconsistent bus potential, contactor jitter, parameter download and action asynchronization, etc. The present application completes one-to-one binding of vehicle and pile by mirror mapping instance, BMS issues a set of limit parameters and occupation, exclusion words; the AGV controller generates a directed cross-bit sequence on the node graph, sets zero flow + zero voltage change + de-bouncing preposition, arranges pre-charging first close, main return second close, and main return first break, pre-charging second break, and downloads CC, CV, and charging power change rate boundaries to the charging power conversion module. Through the technical solutions of the present application, the pre-charging-main loop switching and plug-in processes are constrained by timing and criteria, the power trajectory is predictable, the contactor action is consistent, the parameters and actions are aligned, and the risk of transient impact and false parallel connection is reduced.
[0055] (2) At -10 DEG C and high SOC, the cell polarization increases, the dynamic voltage limit is tightened, and the traditional control is prone to single voltage spikes and early over-limiting. If there is no soft retreat, the current drop is not timely. In the technical solution of the application, the BMS output maps the dynamic voltage limit and the minimum safety margin with temperature, SOC, and SOH, and gives the voltage and current rate boundary; the controller inserts a current limiting feedforward step before enabling the main loop, and tightens the target current and current rate in the constant current stage when the dynamic voltage limit and the minimum safety margin are insufficient; the de-bouncing and zero voltage change check are set in the switching step; the soft retreat is triggered by the rising edge of the plug-in and is linked to the discharging. In the technical solution of the application, the switching and plug-in are constrained by feedforward and hard criteria in the above harsh working conditions, the single spike is suppressed, the current drop and discharging are coordinated, and the action sequence is stable and reproducible.
[0056] (3) When multiple vehicles access the same bus segment, the traditional first-come-first-served and independent charger logic is prone to channel crossing, false parallel connection, and power occupation in adjacent time slots, resulting in unpredictable bus overflows and waiting. The technical solution of the application queues the candidate steps by front, buffer, and resident column partitioning, uses odd-even arrangement and gap rotation to select steps, and uses channel ID and node ID mutual exclusion bitmap for conflict detection; it uses a two-stage commit with reservation-confirmation and time window, constraint ID consistency check, and unselected steps are ordered by aging rules with weighted priority. The application executes the sequence fair, non-crossing, and predictable for the above problems, avoids resource occupation and starvation, improves bus and post utilization, and maintains the determinism and auditability of sequence advancement.
[0057] (4) Field plugging interlocking may be jittery, and if the main loop is directly disconnected or the plug-in intention and electrical access are not distinguished, it is easy to cause flow disconnection, arc light, and bus potential step, affecting subsequent re-handshake and re-precharge. The scheme proposed in the application monitors the rising edge event of plugging, enters the soft exit sequence: first invest in the discharging channel to withdraw the bus energy according to the equivalent admittance curve, while issuing a current drop to zero on the charging side; after meeting the zero flow and zero AV, the main loop is disconnected first and then the precharge is disconnected; after re-handshake, the precharge path is restored by mirror mapping. In the application, the plugging process forms a controlled energy withdrawal and timing bundling, avoiding flow disconnection and potential step, and the conditions for re-handshake and re-precharge are clear, and the process is traceable and reviewable.
[0058] (5) When AGV descends or slows down for regeneration and charging, the regeneration backfill may push the bus voltage up to the CV upper limit, and the traditional control is processed independently, lacking coordination and time slot management with charging and discharging. Figure 4As shown, in the scheme of the application, the AGV controller issues a regeneration control frame to the regeneration power interface according to the mirror instance, obtains power, current, and torque limits by looking up tables according to minimum voltage safety margin, battery temperature, SOC, and SOH, and enables or freezes the odd-even and partition time slots (that is, the AGV controller enters the matching odd-even and partition time slots in the scheduling clock, reads the current step odd-even mark and partition token, sets the channel access and regeneration limit enable bit, downloads the CC / CV and limit parameters, and allows the power channel to start and stop; if it is not in the current time slot, clear the enable bit, freeze the limit table, and keep the last safe output or zero flow instruction, and shield the channel switching request until the next valid time slot); the soft limit sequence is triggered by the zero flow or relay state change; and the charging side parameters and the discharging time window are consistent. For example, when the system clock is 1200 milliseconds, it is in an odd time slot and the partition is in the leading area, the current step odd-even mark is odd and the partition is also in the leading area, the AGV controller sets the channel access and regeneration limit enable bit, issues a constant current target of 80 amperes, a constant voltage target of 738 volts, a current change rate upper limit of 10 amperes per millisecond, a voltage change rate upper limit of 2 volts per millisecond, a regeneration power upper limit of 12 kilowatts, a regeneration current upper limit of 25 amperes, and a resident time slot of 40 milliseconds, and allows the power channel to continue to 1240 milliseconds; when the system clock is 2000 milliseconds, it is in an even time slot and the partition is in the buffer area, the current step mark is odd and the partition is in the leading area, which does not match, the AGV controller clears the channel access and regeneration limit enable bit, freezes the limit table, keeps the last safe output as zero current and issues a zero flow instruction, and shields the channel switching request until the next valid time slot; when the system clock is 3560 milliseconds, it is in an even time slot and the partition is in the resident area, the current step mark is even and the partition is in the resident area, which matches, and the controller receives a state change of the main circuit being ready to be disconnected, enables the regeneration limit and enters the soft limit stage, linearly reduces the regeneration power upper limit from 8 kilowatts to zero according to the set power reduction speed, linearly reduces the regeneration current upper limit to zero according to the set current reduction speed, and freezes the limit at 3600 milliseconds and prohibits channel switching. The application regards charging and discharging, discharging, and regeneration as discrete points and cross-bit steps, sets reachable edges, occupation and forbidden entry conditions, odd-even marks and partition tokens for each step, divides odd or even time slots on the global clock, only allows actions matching the current step label to enter the valid time slot, freezes if it does not match, uses two-stage reservation-confirmation to write to the current step, passes through hard thresholds such as zero flow, zero potential difference, and debounce window before confirmation; uses a mutual exclusion bitmap to resolve conflicts for concurrent resources; puts regeneration and discharging into the same energy boundary, and realizes three-domain linkage through the limit table + soft limit sequence and equivalent admittance switching.And generally with Pack voltage / SOC threshold trigger CC→CV, with fixed hysteresis and simple debouncing; regeneration, charging, and energy discharge are mostly independent, without unified time slots and partition rotation, and lack of zero-flow or zero-potential difference hard access and two-phase commit, and no concurrent scheduling of candidate step freezing-time execution; when multiple vehicles and multiple piles, it is easy to rely on first come first serve, and the transient state of plug-in and backfill lacks trajectory convergence.
[0059] (6) The traditional total energy reading is difficult to locate the energy change of each step, the channel state and the time anchor point, and it is impossible to quickly check and review the abnormal switching, boundary crossing or mismatch. The invention generates a metering segment-channel ID binding with a cross-bit step number as the main index; zero-flow criterion is triggered by dual-channel sampling + zero sequence detection; programmable isolation matrix performs power-on, disconnection and bypass on vehicle and pile channels; metering and isolation frames are attached with mirror number, CRC and timestamp backwriting. The energy increment and isolation state of each step have paragraphized recording and consistency checking, forming an auditable link to support online determination of feasibility and subsequent rearrangement.
[0060] (7) The traditional Pack threshold + CC-CV + fixed hysteresis process does not contain cross-bit sequence and mirror mapping, lacks zero-flow, zero-voltage safety margin access, energy discharge coordination and regeneration time slot control, and concurrent parity, partition rotation and mutual exclusion detection, and the energy transition in the plug-in and switching stage is uncontrollable. The invention uses a directed cross-bit sequence + mirror mapping instance as the skeleton, with pre-occupancy, exclusion screening and zero-flow, zero-voltage safety margin + debouncing threshold, and the charging side parameter download contains a change rate boundary; and with admittance discharge and regeneration limiting based on the minimum voltage safety margin in time slots, supplemented by partition rotation and two-phase commit, paragraphized metering backwriting. At the structural and timing level, the energy channel and concurrent scheduling are uniformly arranged, the switching, plug-in and backfill are predictable, the conflicts and misconnection are controlled, the records and traces are complete, and the engineering requirements of port roll-loading transportation complex working conditions are met.
[0061] From the above specific technical problem solving process can be seen: the technical solution of the application drives the dynamic limit value with the minimum voltage safety margin, tightens the constant current target and the upper limit of constant voltage in real time, and does not let the single cell touch the top first; The hard access / beam threshold of zero current+zero potential difference+debouncing guarantees that pre-charging→main return→CV switching and plug-in only occur within the safety window; Soft current reduction and bus energy discharge cooperate to turn the transient state of switching and plug-in into a controlled trajectory, avoiding the peak to the single cell or bus; Regenerative power is limited by time slot and mirror instance, preventing local overvoltage caused by superimposition of backflow and charging; Vehicle and pile mirroring binding and channel mutual exclusion avoid abnormal charging caused by wrong connection, parallel connection and wrong routing; Subsection metering and event backwriting, abnormality latching and freezing promotion, avoid continuing charging in unknown state. It should be noted that BMS measurement accuracy, temperature and SOC estimation, relay debouncing window, change rate boundary, energy discharge admittance curve and regenerative amplitude limiting table all need to be calibrated according to the specifications of the battery cell and the characteristics of the on-site bus; At the same time, bench / HIL and on-site joint debugging verification should be done, and fault degradation and emergency disconnect path should be set. With these safeguards, the system can better form a three-layer defense of feedforward+access+beam for overcharge causes.
[0062] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.
[0063] Finally: the above is only a preferred embodiment of the present application and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included within the protection scope of the present application.
Claims
1. An AGV charging control system, characterized in that, The system includes a Battery Management System (BMS), AGV controller, relay management module, regenerative power interface, charging power converter, bus de-energizing module, plug-in / plug-out handshake, metering isolation unit, and charging management platform. The BMS provides the AGV controller with cell and module status and voltage / current limits. The AGV controller generates a directed span sequence on the node array based on occupancy markers and prohibition rules, and determines the channel set, parity allocation, partition rotation order, and channel switching constraints. It then sends action commands to the charging power converter and bus de-energizing module, and synchronizes them to the charging management platform. The charging management platform issues a vehicle-charging pile mirror mapping table and number. The plug-in / plug-out handshake completes the process... After mechanical interlocking, identity authentication, insulation and metering isolation are completed, time synchronization is achieved, energy unloading standby is activated, and command arrival is confirmed, the vehicle-end number and the pile-end number are bound to form a mirror mapping instance. The AGV controller issues the cross-position sequence to the charging power converter and the bus unloading module according to the mirror mapping instance, and issues the limiting and switching parameters to the regenerative power interface and the charging power converter. The relay management module controls the opening and closing of the pre-charging circuit and the main circuit according to the confirmation result and the timestamp. The metering isolation unit performs empowerment, isolation and zero-current verification step by step according to the cross-position sequence, and writes back the segmented energy metering, channel identification and timestamp to the AGV controller and the charging management platform.
2. The AGV charging control system according to claim 1, characterized in that, In the AGV controller, the occupancy flag is composed of the contact status of the relay management module, the channel occupancy identifier of the metering isolation unit, the bus potential difference and current zero current flag, and the mechanical interlock and authentication flag of the plug-in / plug-out handshake. The prohibition rules consist of a preset fault isolation table, an insulation incomplete table, an identity inconsistency table, a temperature overrun table, a SOC upper limit table, a voltage upper limit table, and a maintenance lockout table. The node array consists of vehicle-end nodes, pile-end nodes, and bus segment nodes. Vehicle-end nodes include pre-charge circuit nodes, main circuit nodes, regenerative circuit nodes, and Pack terminal nodes. Pile-end nodes include power converter output nodes, metering nodes, isolation nodes, and de-energizing nodes. The channel set consists of pre-charge channels, main charge channels, de-energizing channels, regenerative limiting channels, and their allowed connection edges. The odd-even arrangement is an alternating arrangement of step index and landing point index based on time index. The partition rotation order is the partition rotation order of the leading edge area, buffer zone, and stationary area divided by the yard area and bus segment. The channel switching constraints include zero current criteria, zero potential difference criteria, pre-charge before main circuit closure, soft de-energization before disconnection, relay debounce time window, and cross-channel non-parallel rule set.
3. The AGV charging control system according to claim 2, characterized in that, The BMS performs frame indexing and time stamping on the sampled data of cells and modules, and generates occupancy flag words that include relay contact mirrors, metering isolation mirrors, zero current criteria, and zero bus potential difference criteria; it also generates prohibition rule words based on fault isolation entries, insulation not ready entries, identity inconsistency entries, maintenance lockout entries, temperature window entries, and SOC and voltage upper limit entries. The system calculates a set of limit parameters for cell voltage limiting, module voltage limiting, pack constant voltage target, current limit, and regenerative power limit based on temperature, SOC, aging indicator, charge / discharge rate, and charging stage indicator. It outputs cell indices and sorting vectors for cells with voltage below a preset voltage threshold, temperature above a preset temperature threshold, SOC below 20%, or SoH below 70%. The preset voltage and temperature thresholds are determined based on cell attribute parameters. The cell index represents the cell's position within the battery pack. The sorting vector is generated by calculating a weighted comprehensive score based on the cell's voltage, temperature, SOC, and SoH, and sorting the cells from smallest to largest score. The system also estimates the first and second order rates of change of cell voltage and generates slope and curvature parameters for predictive current limiting. Finally, it packages the occupancy marker, prohibition rule, limit parameter set, cell indices meeting preset threshold requirements, and time stamps into a message and publishes it to the AGV controller along with the vehicle-to-pile mirror mapping instance number.
4. The AGV charging control system according to claim 1, characterized in that, The charging power converter receives charging power conversion action command frames issued by the AGV controller in a directed cross-position sequence. The charging power conversion action command frame includes step number, channel ID, constraint ID, timestamp, CC target current, CV target voltage, charging current change rate and charging voltage change rate limit, and soft deactivation trigger flag. In the pre-charge channel, it performs current limiting power-on according to the limit. In the main charging channel, it drives the CC trajectory according to the table and performs constant voltage switching according to the charging current change rate and charging voltage change rate limit when the switching condition is met. It starts and stops the power channel according to the odd-even arrangement and partition rotation time slot. When it receives the soft deactivation trigger or the zero current criterion of the plug-in node, it reduces the output current to zero according to the zero current sequence and keeps the bus and output terminal potential consistent. It verifies the channel switching constraint according to the constraint ID and performs rejection and status latching for inconsistent commands. It reports the voltage, current, channel status, energy metering and timestamp to the AGV controller and charging management platform according to the segment number.
5. The AGV charging control system according to claim 1, characterized in that, The bus unloading module receives unloading command frames issued by the AGV controller in a directed cross-position sequence. The unloading command frame includes step number, bus segment number, constraint ID, timestamp, target bus potential, equivalent admittance curve, power limit, charging power change rate and charging voltage change rate limit, activation / deactivation threshold, and dwell time slot. It activates electronic load channels or resistor channels in the designated bus segment according to the equivalent admittance curve and forms an unloading trajectory according to the charging power change rate and charging voltage change rate limit. It controls channel activation / deactivation according to odd / even arrangement and partitioned rotation time slots. It executes unloading pre-setup and exit timing according to the plug-in node zero current criterion and metering isolation ready state. It verifies non-crossing and channel switching constraints according to constraint ID and rejects and latches inconsistent commands. It reports the bus potential, unloading current, energy metering, channel identifier, and timestamp to the AGV controller and charging management platform according to segment number.
6. The AGV charging control system according to claim 1, characterized in that, The regenerative power limiting interface receives regenerative control frames issued by the AGV controller based on a mirror mapping instance. These frames include a step number, channel identifier, constraint identifier, timestamp, and indexes for a regenerative power upper limit table, DC bus voltage upper limit, braking torque limiting table, power and current ramp-up slope, short-time suppression threshold, and dwell time slot, indexed by minimum safety margin, temperature, state of charge, and health status. Based on the channel identifier mapping, it generates motor braking torque and DC-side current limits, and outputs them through rate limiting and feedforward filtering shaping. Limiting is enabled or frozen in the corresponding time slot according to odd / even alignment and partitioned rotation. Upon receiving a zero-current criterion for plug-in / plug-out nodes or a status change in the relay management module, it enters the soft limiting stage and executes cancellation or restoration according to the timing table. It verifies non-crossing and channel switching constraints according to the constraint identifier and rejects and latches inconsistent commands. Finally, it reports the regenerative power, DC current, motor braking torque, channel identifier, and timestamp to the AGV controller and charging management platform according to segment numbers.
7. The AGV charging control system according to claim 1, characterized in that, The metering isolation unit uses the step number of the cross-position sequence as the main index to establish the binding relationship between the segmented metering channels and the channel identifiers. It uses a combination of dual-channel current sampling and zero-sequence detection to complete the zero-current criterion triggering. It performs three-state switching of power-on, disconnection and bypass for the vehicle-end and pile-end channels according to the programmable isolation matrix. It performs numbering and cyclic redundancy verification on the metering frames and isolation frames based on the mirror mapping instance and writes the timestamp. It sets the hierarchical isolation priority and interlocking conditions for the bus section and branch section. It coordinates the opening and closing sequence of the relay management module through edge-triggered zero-current interlocking and potential difference interlocking. It completes the segmented energy metering accumulation and metering segment number update in steps, and writes back the metering data, isolation status word, channel identifier and timestamp to the AGV controller and charging management platform in the form of messages.
8. An AGV charging control method, employing an AGV charging control system as described in any one of claims 1-7, characterized in that, The method includes: Step 1, Establish Mapping and Baseline: Obtain the vehicle-end and pile-end numbers of the AGV to generate a mirror mapping instance, complete the initialization of time base, serial number and occupancy code, collect the individual cell voltage, module-end voltage and total pack voltage, DC bus current, charging output current and regeneration return current, cell temperature, module temperature and pack shell temperature, pre-charge contact and main contact status and plug-in interlock status, and obtain the bus potential difference and zero current flag to form occupancy mark word and prohibition rule word; Step 2, Sequence Arrangement and Round-Robin: Generate a directed spanning sequence on the discrete node set according to the occupancy mark and the prohibition rule, determine the allowed connection edges, step order and parity arrangement, and load the candidate steps into the partition queue according to the region and the bus line segment, and select the current execution step through the round-robin strategy; Step 3, Constraint Verification and Channel Admission: Perform rule matching and state latching on non-crossing, zero-current, zero-potential difference, debouncing time window and channel switching sequence; enable or isolate the target channel and perform zero-current verification; record channel identifier and metering segment. Step 4, Charging power parameterization and energy removal coordination: Generate charging constant current target, charging constant voltage target, charging power change rate boundary, charging limit parameter and dwell time slot for the current step and complete parameter download. When charging energy withdrawal is involved, generate equivalent charging admittance curve, charging power and voltage change rate boundary, charging start and stop threshold and dwell time slot simultaneously. Step 5, Regeneration Limiting and Time Slot Control: Based on the minimum safety margin of a single cell during charging, temperature and state of charge index, calculate the limiting of regeneration power, DC current and motor torque, enable or freeze the limiting according to odd and even time slots and partition time slots, and enter the soft limiting stage to perform limiting cancellation or restoration when plugging / unplugging zero current or relay management status changes. Step 6, Execution Confirmation, Advancement and Constraint Archiving: Confirm zero current, zero potential difference, parameter readiness and consistency and write timestamps and event sequences. If the conditions are met, advance to the next step. If not, freeze and rearrange the path and parameters. After reaching the termination node, execute soft shutdown, channel withdrawal and disconnection in sequence, and complete metering write-back and record archiving.
9. The AGV charging control method according to claim 8, characterized in that, In step 2, the sequence arrangement includes: representing the discrete node set as a directed graph G(V,E), with V consisting of node type, bus segment number, and region identifier, and E consisting of allowed connecting edges and their channel types; filtering E using occupancy tags and prohibition rules to obtain E'; adding a step order index, parity tag, channel ID, and timestamp window to each edge; using an adjacency list and bitmap to lock the occupancy status of nodes and edges; using the starting node set S sorted lexicographically and stably by timestamp as the root, generating a directed spanning sequence on the subgraph (V,E') by restricted topology traversal, performing precondition checks, edge label consistency checks, and time window consistency checks at each step during the traversal, and writing it into the sequence using a two-stage reservation and commit process.
10. The AGV charging control method according to claim 8, characterized in that, In step 2, the round-robin selection includes: writing candidate steps into the partition queue {Q} by region and bus segment. K Each queue records its queue token, remaining weight, and last service time; a gap-based round-robin algorithm is used to process {Q}. K The selection process involves defining the parity of eligible steps based on a parity mapping table and performing conflict detection using a mutual exclusion bitmap of channel IDs and node IDs. For candidate steps that pass the detection, a two-phase submission (reservation-confirmation) is performed. The reservation phase sets lock flags for nodes and edges, while the confirmation phase completes the submission based on timestamp windows and constraint IDs. In the event of concurrent contention, a decision is made based on the {priority, timestamp, sequence number} triple. Unselected candidate steps have their weights updated according to aging rules, and their enqueue order is preserved for the next round of selection.
Citation Information
Patent Citations
Intelligent transportation system based on Internet of Things and application
CN106354142A
Smart city energy dynamic scheduling system and method based on big data analysis
CN120672078A