A static start frequency converter commutation control system and method

Abstract

The application discloses a static starting frequency converter commutation control system and method, and relates to the technical field of static starting frequency converter commutation control. The system constructs a current state set according to a machine-side line voltage, a machine-side phase current, a DC reactor current, an excitation current and a machine-side inverter bridge thyristor conduction state; generates a one-time effective zero-crossing identification according to a machine-side line voltage zero-crossing event; generates a candidate commutation control scheme by using a dynamic conduction conflict relationship graph; calculates a reachable state interval sequence based on a timing mixed state model, generates commutation permission information when meeting a safe state domain and a target current path state domain constraint, and implements monitoring and safe rollback according to an actual state trajectory. The application solves the problems of state uncertainty, zero-crossing misjudgment and commutation safety shortage in the low-speed starting stage, can reduce mis-triggering, commutation failure and current impact, and improves the safety, robustness and continuous operation capability of the starting process.

Description

Technical Field

[0001] This invention relates to the field of commutation control technology for static start frequency converters, and more particularly to a commutation control system and method for static start frequency converters. Background Technology

[0002] Static start inverters typically use coordinated control of grid-side rectification, DC link, and machine-side inverter to gradually drive a synchronous motor from a standstill to its self-sustaining speed. The machine-side inverter bridge often uses thyristor devices, and its commutation process requires consideration of machine-side voltage, current, rotor position, DC current, and thyristor turn-off recovery state to determine the triggering timing. Existing control methods usually rely on instantaneous sampled values, machine-side line voltage zero-crossing signals, and preset commutation sequences for logical judgment, outputting corresponding trigger pulses after a set threshold is met.

[0003] During the stationary and low-speed start-up phases of a synchronous motor, the amplitude of the induced voltage on the motor side is low, and the zero-crossing signal is easily affected by harmonics, commutation gaps, measurement noise, and sampling delays. Simultaneously, the thyristor turn-off recovery time, current transfer process, and motor parameters also change with operating conditions. Therefore, commutation judgments based on a single estimated state or fixed threshold may not accurately reflect the actual operating state of the equipment, and it is particularly difficult to determine whether all current commutation conditions are valid within the state error range.

[0004] The aforementioned deviations in state judgment may lead to misidentification of zero-crossing events, improper triggering timing, premature re-triggering of thyristors, or failure to establish the target current path, which in turn can cause commutation failure, abnormal DC current, machine-side current surges, and electromagnetic torque fluctuations. In severe cases, it may even force the interruption of the static start-up process, increase the electrical stress on the thyristors and synchronous motor windings, and affect the safety and reliability of the gas turbine generator set's start-up process. Summary of the Invention

[0005] To address the problem of safe commutation control under low-speed uncertain conditions, this invention proposes a commutation control system and method for a static start frequency converter.

[0006] The present invention achieves the above objectives through the following technical solutions:

[0007] A commutation control system for a static start frequency converter, comprising:

[0008] The state construction module is used to construct a current state set, including rotor phase reference and defined by state quantity error boundaries, based on the machine-side line voltage, machine-side phase current, DC reactor current, excitation current and the conduction state of the machine-side inverter bridge thyristors.

[0009] The event authentication module is used to generate a one-time valid zero-crossing identifier based on the phase sequence, time, and angular consistency with the rotor phase reference of the zero-crossing event of the machine-side line voltage.

[0010] The scheme generation module is used to screen candidate conduction combinations based on the one-time effective zero-crossing identifier, the current state set, and the turn-off recovery state of the inverter bridge thyristors on the machine side, using a dynamic conduction conflict relationship diagram, and to combine the candidate conduction combinations, target current paths, and rectifier, inverter, and excitation coordinated actions into candidate commutation control schemes.

[0011] The safety verification module is used to calculate the sequence of reachable state intervals of the current state set under each candidate commutation control scheme based on the timing hybrid state model. When each reachable state interval in the sequence is located in the safe state domain and the last reachable state interval is located in the target current path state domain, it generates a binding one-time valid zero-crossing identifier and commutation permission information of the candidate commutation control scheme.

[0012] The execution and rollback module is used to execute the bound candidate commutation control scheme when both the one-time valid zero-crossing flag and commutation permission information are valid and unused; when the actual state trajectory exceeds the reachable state interval sequence or fails to enter the target current path state domain within the preset confirmation time window, it performs a safety rollback operation, including locking the inverter bridge thyristor trigger pulse on the machine side, reducing the DC reactor current to a safe threshold, and rolling back to the commutation control state when the most recent commutation was successful.

[0013] Preferably, the state construction module includes:

[0014] The state estimation unit is used to determine the estimated values ​​of basic state variables, including rotor phase reference and rotor angular velocity, based on the machine-side line voltage, machine-side phase current, DC reactor current and excitation current.

[0015] The conduction topology verification unit is used to determine the expected conduction topology of the machine-side inverter bridge thyristors based on the triggering records of the machine-side inverter bridge thyristors, determine the observed current path based on the current continuity relationship between the machine-side phase current and the DC reactor current, and determine the set of feasible machine-side inverter bridge thyristor conduction topologies based on the compatibility between the expected conduction topology of the machine-side inverter bridge thyristors and the observed current path, as well as the unidirectional conduction constraints, bridge arm conduction constraints and current continuity constraints of the machine-side inverter bridge thyristors.

[0016] The state set generation unit is used to determine the corresponding state quantity error boundary based on the basic state quantity estimate for each feasible machine-side inverter bridge thyristor conduction topology, and to constrain and correct the state range defined by the state quantity error boundary according to the current continuity relationship between the machine-side phase current and the DC reactor current, the phase relationship between the machine-side line voltage and the rotor phase reference, and the allowable conversion relationship of the machine-side inverter bridge thyristor conduction topology in adjacent control cycles, thereby generating a sub-state set corresponding to each feasible thyristor conduction topology, and determining the union of each sub-state set as the current state set.

[0017] Preferably, the event authentication module includes:

[0018] The event matching unit is used to generate a predicted zero-crossing window for each sub-state set in the current state set based on the corresponding feasible machine-side inverter bridge thyristor conduction topology, rotor phase reference, rotor angular velocity and preset zero-crossing phase sequence, and determine the event matching state set based on the matching results between the machine-side line voltage zero-crossing event and each predicted zero-crossing window.

[0019] The topology verification unit is used to determine the event matching state set as the authentication state set; when there is a previously generated one-time valid zero-crossing identifier, it determines the set of thyristor conduction topologies that can reach the machine-side inverter bridge based on the current state set bound to the previously generated one-time valid zero-crossing identifier and the preset allowed commutation sequence, and filters the authentication state set accordingly.

[0020] The common conduction combination determination unit is used to determine the event-compatible conduction combination set supported by each sub-state set in the authentication state set in all states, and to determine the intersection of each event-compatible conduction combination set as the common event-compatible conduction combination set.

[0021] The identifier generation unit is used to update the current state set with the authentication state set when both the authentication state set and the common event compatible conduction combination set are not empty, and to generate a zero-crossing event of the bonding machine side line voltage, the updated current state set, the common event compatible conduction combination set, the identifier valid time window, and a one-time valid zero-crossing identifier of the identifier usage status.

[0022] Preferably, the scheme generation module includes:

[0023] The conflict relationship construction unit is used to set the relationship between the machine-side inverter bridge thyristor nodes in the same bridge arm as a fixed conflict relationship for each sub-state set in the current state set, with each machine-side inverter bridge thyristor as a node, and to set a dynamic conduction conflict relationship according to the corresponding feasible machine-side inverter bridge thyristor conduction topology, machine-side inverter bridge thyristor turn-off recovery state, DC reactor current variation range, and time interval between adjacent machine-side inverter bridge thyristor trigger pulses, and to construct a corresponding dynamic conduction conflict relationship diagram.

[0024] The combination filtering unit is used to exclude, from the set of common event compatible conduction combinations bound by the one-time valid zero-crossing identifier, conduction combinations that contain conflicting node pairs in any dynamic conduction conflict relationship graph, and determine the remaining conduction combinations as the candidate conduction combinations;

[0025] The scheme arrangement unit is used to determine, for each candidate conduction combination, the common execution time window in which each sub-state set in the current state set satisfies the turn-off recovery condition, current transfer condition, and excitation support condition of the inverter bridge on the machine side, and to determine the target current path corresponding to each candidate conduction combination. The candidate conduction combination, the rectification, inversion and excitation coordinated action within the common execution time window, the target current path and the preset confirmation time window are combined into the candidate commutation control scheme.

[0026] Preferably, the security verification module includes:

[0027] The branch construction unit is used to generate state evolution branches based on the timing hybrid state model, according to the discrete mode conversion sequence within the common execution time window and the turn-off recovery time interval of the machine-side inverter bridge thyristors, for each candidate commutation control scheme and each sub-state set in the current state set. The discrete modes of the timing hybrid state model include current conduction, rectification current reduction, machine-side inverter bridge thyristor turn-off recovery, current transfer, target conduction, and commutation anomaly mode. Each state evolution branch is bound to the corresponding sub-state set, discrete mode conversion sequence, and machine-side inverter bridge thyristor turn-off recovery time interval.

[0028] The bidirectional reachability calculation unit is used to calculate the reachable state interval sequence in the forward direction along each state evolution branch, and to calculate the reverse reachable state interval of each state evolution branch at each prediction time in the reverse direction from the target current path state domain according to the remaining rectification, inversion and excitation coordinated action of the corresponding candidate commutation control scheme after each prediction time.

[0029] The permit generation unit is used to determine the commutation permit information generation condition as follows: the reachable state intervals of each state evolution branch at each prediction time are all located within the intersection of the safe state domain and the corresponding reverse reachable state interval, and the terminal reachable state intervals are all located within the target current path state domain. When the commutation permit information generation condition is met, commutation permit information is generated that binds the corresponding candidate commutation control scheme, the one-time effective zero-crossing identifier, and the set of state evolution branches.

[0030] Preferably, the commutation license information further includes a license validity period window and license status;

[0031] The execution and rollback module includes a permission verification unit, which is used to call the state construction module before execution to generate a real-time state set for execution entry verification, and determine the real-time machine-side inverter bridge thyristor conduction topology based on the thyristor conduction status of the machine-side inverter bridge collected before execution; determine whether there is a common state evolution branch from the state evolution branch set bound to the commutation permission information, so that the real-time state set is included in the sub-state set bound to the state evolution branch, and the real-time machine-side inverter bridge thyristor conduction topology is consistent with the feasible machine-side inverter bridge thyristor conduction topology corresponding to the sub-state set; when there is a common state evolution branch, the one-time valid zero-crossing identifier bound to the commutation permission information is within the identifier valid time window and the identifier usage status is unused, and the commutation permission information is within the permission valid time window and the permission status is unused, an execution permission signal is generated for the corresponding candidate commutation control scheme.

[0032] Preferably, the execution and rollback module further includes a scheme optimization unit;

[0033] The scheme selection unit is used to normalize the distance from the reachable state interval of each state evolution branch to the boundary of the safe state domain and the distance to the boundary of the corresponding reverse reachable state interval at each prediction time for each candidate commutation control scheme that has generated an execution permission signal, according to the preset state scale corresponding to each state variable. The minimum value of the two types of normalized distances in all state evolution branches and all prediction times is determined as the scheme robustness margin, and the candidate commutation control scheme with the largest scheme robustness margin is determined as the target commutation control scheme.

[0034] The execution and rollback module executes within the common execution time window corresponding to the target commutation control scheme, and after the output side inverter bridge thyristor trigger pulse, updates the permission status and usage status corresponding to the target commutation control scheme to the used status respectively.

[0035] Preferably, the execution and rollback module further includes:

[0036] The trajectory reconstruction unit is used to call the state construction module at each sampling time to generate a real-time state set after the target commutation control scheme is executed, and to determine the real-time machine-side inverter bridge thyristor conduction topology based on the machine-side inverter bridge thyristor conduction state at each sampling time, thereby forming the actual state trajectory.

[0037] The branch monitoring unit is used to determine the set of state evolution branches corresponding to the target commutation control scheme as the current retained set at the first sampling time. At each sampling time, it determines the set of topologically consistent state evolution branches that are consistent with the real-time thyristor conduction topology of the machine-side inverter bridge from the current retained set, and determines the union of the reachable state intervals of the set of topologically consistent state evolution branches at the corresponding prediction time as the union of topologically consistent reachable state intervals. When the real-time state set is included in the union of topologically consistent reachable state intervals, the topologically consistent state evolution branch that intersects with the reachable state interval and the real-time state set is determined as the current retained set at the next sampling time.

[0038] The path confirmation unit is used to determine that the actual state trajectory enters the target current path state domain when the real-time state set is included in the target current path state domain and the real-time machine-side inverter bridge thyristor conduction topology is consistent with the machine-side inverter bridge thyristor conduction topology corresponding to the candidate conduction combination within a preset confirmation time window.

[0039] The rollback control unit is used to execute the safe rollback operation when the set of evolutionary branches of the topology-consistent state is empty, the real-time state set is not included in the union of the topology-consistent reachable state intervals, or the confirmation conditions of the path confirmation unit are not met.

[0040] Preferably, the execution and rollback module further includes:

[0041] An error boundary correction unit is used to determine the state variables and deviation directions that were not fully covered initially after the rollback control unit performs a safe rollback operation due to the non-empty set of topologically consistent state evolution branches and the real-time state set not being included in the union of topologically consistent reachable state intervals. Under the condition that the error boundaries of other state variables remain unchanged, it determines the minimum boundary correction amount required to ensure that the recalculated union of topologically consistent reachable state intervals fully covers each formed real-time state set, and sends the minimum boundary correction amount to the state construction module. The state construction module corrects the corresponding state variable error boundaries according to the minimum boundary correction amount and updates the current state set. The permission verification unit updates the permission status of the phase commutation permission information generated based on the current state set before the update to an invalid status.

[0042] The scheme isolation unit is used to generate scheme feature identifiers based on the candidate conduction combinations, target current paths, and the type and order of rectification, inversion, and excitation coordinated actions in the target commutation control scheme. The safety verification module performs non-executable re-verification of candidate commutation control schemes with scheme feature identifiers based on the updated current state set. If a candidate commutation control scheme still does not meet the conditions for generating commutation permission information, the scheme feature identifier is added to the scheme isolation set, prohibiting the candidate commutation control scheme with the scheme feature identifier from participating in the subsequent generation of commutation permission information and determination of the target commutation control scheme.

[0043] A commutation control method for a static start frequency converter includes:

[0044] Based on the machine-side line voltage, machine-side phase current, DC reactor current, excitation current, and the conduction state of the machine-side inverter bridge thyristors, a current state set is constructed, including the rotor phase reference and defined by the state quantity error boundary.

[0045] A one-time valid zero-crossing identifier is generated based on the phase sequence, time, and angle consistency with the rotor phase reference of the zero-crossing event of the machine-side line voltage.

[0046] Based on the one-time effective zero-crossing identifier, the current state set, and the turn-off recovery state of the inverter bridge thyristors on the machine side, candidate conduction combinations are screened using a dynamic conduction conflict relationship diagram, and the candidate conduction combinations, target current paths, and rectifier, inverter, and excitation coordinated actions are combined into candidate commutation control schemes.

[0047] Based on the timing hybrid state model, the sequence of reachable state intervals of the current state set under each candidate commutation control scheme is calculated. When each reachable state interval in the sequence is located in the safe state domain and the last reachable state interval is located in the target current path state domain, a binding one-time valid zero-crossing identifier and the corresponding candidate commutation control scheme commutation permission information are generated.

[0048] When both the one-time valid zero-crossing identifier and the commutation permission information are valid and unused, the candidate commutation control scheme bound to the commutation permission information is executed; when the actual state trajectory exceeds the corresponding reachable state interval sequence, or when the target current path state domain is not entered within the preset confirmation time window, the thyristor trigger pulse of the blocking machine-side inverter bridge is used to reduce the DC reactor current to a safe threshold and return to the commutation control state when the most recent commutation was successful.

[0049] The beneficial effects of this invention are as follows: A current state set, defined by state quantity error boundaries, is constructed based on the generator-side voltage, current, excitation state, and the thyristor conduction state of the generator-side inverter bridge. Phase sequence, time, and angle consistency verification is performed on generator-side line voltage zero-crossing events. Candidate commutation control schemes are screened by combining thyristor turn-off recovery states and dynamic conduction conflict relationships. This enables consistency verification of commutation events and candidate conduction combinations even with measurement errors, parameter deviations, and conduction topology uncertainties, reducing the impact of erroneous zero-crossing signals and local state judgments on commutation control. A timed hybrid state model is used to verify the reachability of each candidate commutation control scheme. Commutation permission information is generated only when the state evolution process continuously remains in the safe state domain and can enter the target current path state domain. Monitoring and safe backoff are implemented during execution based on the actual state trajectory. This method improves the reliability of commutation triggering conditions, reduces thyristor false triggering, commutation failure, current surges, and startup interruptions, and enhances the safety, robustness, and continuous operation capability of static start-up inverters under complex startup conditions. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the modular structure of the system according to an embodiment of the present invention; Figure 2 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0052] like Figure 1 As shown, this is an embodiment of the present invention, which provides a commutation control system for a static start frequency converter, including a state construction module, an event authentication module, a scheme generation module, a safety verification module, and an execution and rollback module.

[0053] Taking a six-pulse thyristor-type static starter inverter driving a synchronous motor as an example, the static starter inverter includes a grid-side rectifier bridge, a DC reactor, a machine-side inverter bridge, and an excitation device. The machine-side inverter bridge includes the thyristor in the upper arm of phase A. Phase A lower bridge arm thyristor Phase B upper arm thyristor B+, Phase B lower arm thyristor C-phase upper arm thyristor and C-phase lower arm thyristor .

[0054] Unless otherwise specified, voltage, current, time, angle, angular velocity, resistance, and inductance are expressed in volts, amperes, seconds, radians, radians per second, ohms, and henries, respectively. Variables marked with "^" in the formulas represent estimated values; superscripts "..." "and" "" represents the lower bound direction and the upper bound direction of the interval respectively, and does not represent positive or negative operations.

[0055] The controller collects the machine-side line voltage, machine-side phase current, DC reactor current, excitation current, and the triggering and conduction feedback of the machine-side inverter bridge thyristors. All sampled signals are processed according to a unified time base. The controller can consist of a processor, programmable logic devices (PLDs), and a trigger interface. The processor performs state estimation, event authentication, commutation scheme generation, and safety verification. The PLDs perform zero-crossing event capture, trigger pulse output, pulse readback, and emergency interlocking.

[0056] Specifically, the controller acquires the machine-side line voltage, machine-side phase current, DC reactor current, excitation current, and the conduction status of the machine-side inverter bridge thyristors according to the control cycle. The state construction module determines the possible current operating states of the synchronous motor and the static start inverter based on each sampled quantity, and considers sensor errors, sampling delays, filter phase shifts, and equipment parameter deviations to form a current state set defined by the error boundaries of the state quantities. The current state set includes at least the rotor phase reference, and also retains the uncertainty range of continuous states and possible machine-side inverter bridge thyristor conduction topologies.

[0057] In one specific embodiment, the state construction module includes a state estimation unit, a conduction topology verification unit, and a state set generation unit.

[0058] The state estimation unit is used to determine the estimated values ​​of basic state variables, including rotor phase reference and rotor angular velocity, based on the machine-side line voltage, machine-side phase current, DC reactor current, and excitation current. Specifically, the state estimation unit performs zero-bias correction, range conversion, filtering, and time alignment on the machine-side line voltage, machine-side phase current, DC reactor current, and excitation current. The machine-side three-phase voltage and three-phase current can be converted into voltage components in a stationary two-phase coordinate system through conventional coordinate transformation. and current components After considering the stator resistance of the synchronous motor, the estimated stator flux linkage is updated according to the following formula: ; Indicates the first Each sampling time at The estimated stator flux linkage on the coordinate axes, in Weber dimension; For coordinate axis labels; This represents the estimated value of the stator flux linkage on the corresponding coordinate axis at the previous sampling time; This indicates the sampling period, measured in seconds. Indicates the first Each sampling time at Machine-side voltage components on the coordinate axes; This represents the corresponding machine-side current component, with dimensions in amperes; This represents the equivalent resistance of each phase of the stator in a synchronous motor.

[0059] The rotor phase reference can be expressed as: ; Indicates the first The rotor phase reference at each sampling moment is expressed in electrical angles, with the dimension in radians; This represents the four-quadrant arctangent function that determines the phase angle based on two coordinate components; and These represent the stator flux linkage in shaft and Estimated value on the axis; It represents the total phase compensation amount caused by the wiring relationship of the voltage measurement channel, the phase shift of digital filtering, the coordinate axis direction, and the installation deviation, with the dimension in radians.

[0060] The rotor angular velocity is determined based on the difference in rotor phase reference between adjacent sampling times and the sampling period. When calculating the phase difference, the phase difference is factored into one electrical angle period to avoid abrupt changes in angular velocity when the phase crosses the period boundary.

[0061] When the synchronous motor is stationary and the machine-side induced voltage is below the phase estimation threshold, the rotor phase reference is initialized using probe pulses. After the excitation current reaches a preset probe value, the machine-side inverter bridge sequentially outputs short-time probe conduction combinations that do not form a direct connection between bridge arms. The machine-side line voltage response or DC reactor current change under each probe conduction combination is collected and compared with the rotor phase partition table. The rotor phase partition table can be established through synchronous motor parameter models, shutdown tests, or real-time digital simulations, recording the line voltage response polarity, amplitude relationship, or DC reactor current change relationship corresponding to different rotor electrical angle intervals. The state estimation unit determines the electrical angle interval of the rotor based on the matching result between the response quantity and the rotor phase partition table, uses the center value of this interval as the initial rotor phase reference, and includes the interval half-width, measurement error, and probe response error in the state quantity error boundary of the rotor phase reference.

[0062] The duration and upper current limit of the probe pulse are set to ensure that they do not generate continuous electromagnetic torque and do not exceed the allowable current of the thyristors and DC reactors in the machine-side inverter bridge. An interval of not less than the turn-off recovery time of the machine-side inverter bridge thyristors is reserved between adjacent probe pulses. Once the machine-side induced voltage reaches the phase estimation threshold, the state estimation unit switches to continuous phase estimation based on the machine-side line voltage and stator flux linkage. When a synchronous motor is equipped with a rotor position sensor, the rotor phase reference can also be initialized using the sensor's measurements, and the sensor's measurement accuracy and installation deviation can be included in the corresponding state quantity error boundaries.

[0063] The conduction topology verification unit is used to determine the expected conduction topology of the machine-side inverter bridge thyristors based on the triggering records of the machine-side inverter bridge thyristors, determine the observed current path based on the current continuity relationship between the machine-side phase current and the DC reactor current, and determine the set of feasible machine-side inverter bridge thyristor conduction topologies based on the compatibility between the expected conduction topology and the observed current path, as well as the unidirectional conduction constraints, bridge arm conduction constraints and current continuity constraints of the machine-side inverter bridge thyristors.

[0064] Specifically, the conduction topology verification unit stores the trigger records of the thyristors in the machine-side inverter bridge, including at least the thyristor number, trigger time, pulse width, and pulse readback result. For example, the most recent valid trigger record corresponds to... and And when no other valid commutation pulses have been output, the expected topology of the inverter bridge thyristors on the machine side is as follows: If already submitted to and Output commutation pulse, but If the current in the branch has not been completely transferred, then... and All of these are considered as expected machine-side inverter bridge thyristor conduction topologies to be verified.

[0065] The observed current path is determined based on the current continuity relationship between the machine-side phase current and the DC reactor current. Taking phase A flowing in and phase B flowing out as an example, when the following conditions are met: At that time, the observed current path is compatible with the A+B− conducting topology. , and These represent the currents on the camera side of phases A, B, and C, respectively. This refers to the current in the DC reactor. This represents the maximum permissible deviation between the phase A current and the DC reactor current, measured in amperes. This represents the maximum permissible deviation between the reverse current of B and the DC reactor current, measured in amperes. This represents the zero-current threshold for the non-conducting phase, with dimensions in amperes; symbol This represents the absolute value. The above deviations and thresholds cover sensor measurement error, sampling quantization error, and allowable leakage current.

[0066] The remaining observation current paths are determined according to the positive and negative directions of the corresponding phase current and the continuity of the DC reactor current.

[0067] The conduction topology verification unit verifies the conduction topologies of each expected machine-side inverter bridge thyristor. The retained conduction topologies simultaneously meet the following conditions: they are interpretable from trigger records; they are compatible with the observed current path; they comply with the unidirectional conduction constraint of the machine-side inverter bridge thyristors; they do not form a direct path where thyristors on the same bridge arm conduct simultaneously; and they comply with the current continuity constraint that the DC reactor current cannot change instantaneously. When the sampled data is compatible with two commutation transition topologies simultaneously, both topologies are retained to form a set of feasible machine-side inverter bridge thyristor conduction topologies.

[0068] The state set generation unit is used to determine the corresponding state quantity error boundary based on the basic state quantity estimate for each feasible machine-side inverter bridge thyristor conduction topology. It also performs constraint correction on the state range defined by the state quantity error boundary according to the current continuity relationship between the machine-side phase current and the DC reactor current, the phase relationship between the machine-side line voltage and the rotor phase reference, and the allowable transition relationship of the machine-side inverter bridge thyristor conduction topology in adjacent control cycles. The unit generates a sub-state set corresponding to each feasible machine-side inverter bridge thyristor conduction topology and determines the union of each sub-state set as the current state set.

[0069] Specifically, continuous state quantities include machine-side line voltage, machine-side phase current, DC reactor current, excitation current, rotor phase reference, and rotor angular velocity. The initial state range of each state variable is represented as follows: ; Number the state variables; For the first The actual value of each state variable; For the first Estimates of the state variables; For the first The lower deviation boundary of each state quantity; For the first Upper deviation boundary of each state quantity; , and They have the same dimensions.

[0070] The state quantity error boundary is determined based on sensor accuracy, zero-bias residual, filter phase shift, sampling time error, synchronous motor parameter error, and calculation rounding error.

[0071] The state set generation unit intersects the initial state range with the corresponding constraints for each feasible machine-side inverter bridge thyristor conduction topology. Constraints include the three-phase current balance relationship, the current continuity relationship between the machine-side phase current and the DC reactor current, the phase relationship between the machine-side line voltage and the rotor phase reference, and the permissible transition relationships of the machine-side inverter bridge thyristor conduction topology in adjacent control cycles. For example, in the A+B− topology, phase A current should be in the positive current range, phase B current should be in the negative current range, and phase C current should be in the zero current error range. If the state range corresponding to a certain conduction topology is an empty set after constraint correction, that conduction topology is deleted. Each retained feasible machine-side inverter bridge thyristor conduction topology corresponds to a sub-state set, and the union of all sub-state sets constitutes the current state set.

[0072] When the machine-side line voltage crosses zero, the event authentication module verifies the zero-crossing channel, zero-crossing direction, zero-crossing time, preset zero-crossing phase sequence, and the angular relationship between the zero-crossing event and the rotor phase reference. Upon successful authentication, a one-time valid zero-crossing identifier is generated. This identifier has a valid time window and usage status, corresponding only to the current zero-crossing event and the current commutation process.

[0073] In one specific embodiment, the event authentication module includes an event matching unit, a topology verification unit, a common conduction combination determination unit, and an identifier generation unit.

[0074] The event matching unit is used to generate predicted zero-crossing windows for each sub-state set in the current state set, based on the corresponding feasible machine-side inverter bridge thyristor conduction topology, rotor phase reference, rotor angular velocity, and preset zero-crossing phase sequence. It then determines the event matching state set based on the matching results between the machine-side line voltage zero-crossing event and each predicted zero-crossing window. Specifically, the machine-side line voltage undergoes low-pass filtering and hysteresis comparison before zero-crossing detection. Taking positive zero-crossing as an example, the machine-side line voltage enters a waiting-to-trigger state after falling below the negative hysteresis threshold. Subsequently, a positive zero-crossing event is generated when the voltage first reaches or exceeds zero, and it is latched until the machine-side line voltage exceeds the positive hysteresis threshold. Negative zero-crossing is judged in the opposite direction. Through hysteresis and latching processing, repetitive zero-crossing events caused by commutation gaps and harmonics are suppressed. Each machine-side line voltage zero-crossing event records at least the line voltage path, zero-crossing direction, and zero-crossing timestamp.

[0075] The event matching unit, for each sub-state set in the current state set, calculates the possible time and angle ranges of the next machine-side line voltage zero-crossing event based on the rotor phase reference interval, rotor angular velocity interval, preset zero-crossing phase sequence, and corresponding feasible machine-side inverter bridge thyristor conduction topology, forming a predicted zero-crossing window. The rotor mechanical angle and rotor electrical angle satisfy the following relationship: ; The rotor electrical angle; For rotor mechanical angle; denoted as the number of pole pairs of the synchronous motor, and denoted as a dimensionless positive integer.

[0076] The rotor electrical angles corresponding to different machine-side line voltage channels, zero-crossing directions, and rotation directions are pre-stored in a zero-crossing phase mapping table. This table also considers the phase compensation caused by voltage measurement wiring phase, digital filtering delay, and rotor magnetic pole axis deviation. When the rotor angular velocity is sufficient to form a stable time prediction, the predicted zero-crossing time range is calculated based on the angle difference between the next zero-crossing phase and the current rotor phase reference, as well as the rotor angular velocity range. When the synchronous motor is in a very low-speed phase, it is directly determined whether the rotor phase reference range at the time of the zero-crossing event intersects with the corresponding zero-crossing phase range. When the channel, direction, and preset zero-crossing phase sequence of the machine-side line voltage zero-crossing event are consistent, the zero-crossing time is within the predicted zero-crossing time range, and the rotor phase corresponding to the zero-crossing event is within the predicted angle range, the zero-crossing event matches the corresponding sub-state set. All matched sub-state sets constitute the event matching state set.

[0077] The verification unit is used to determine the event matching state set as the authentication state set. When a previously generated one-time valid zero-crossing flag exists, it determines the set of reachable machine-side inverter bridge thyristor conduction topologies based on the current state set bound to the previously generated one-time valid zero-crossing flag and the preset allowed commutation sequence, and filters the authentication state set accordingly. Specifically, if a previously generated one-time valid zero-crossing flag does not exist after system startup, the event matching state set is directly determined as the authentication state set. When a previously generated one-time valid zero-crossing flag exists, the current state set bound to the previously generated one-time valid zero-crossing flag is read, and the feasible machine-side inverter bridge thyristor conduction topologies that the current machine-side line voltage zero-crossing event may reach are determined based on the feasible machine-side inverter bridge thyristor conduction topologies and the preset allowed commutation sequence in the current state set.

[0078] In the forward rotation case, a preset allowed commutation sequence is:

[0079] ;

[0080] In the formula, the arrow indicates the allowed direction of topology switching of the thyristor conduction between adjacent commutation steps on the machine side of the inverter bridge; , and These represent the upper bridge arm thyristors of phases A, B, and C, respectively. , and These represent the corresponding lower bridge arm thyristors. During reverse operation, the reverse order corresponding to the rotation direction is used. The set of accessible inverter bridge thyristor conduction topologies is obtained by expanding according to the starting topology corresponding to the previous valid zero-crossing identifier, the number of zero-crossing events that have occurred, and the allowed commutation sequence. In the certified state set, sub-state sets whose corresponding topologies do not belong to this accessible inverter bridge thyristor conduction topology set are deleted.

[0081] The common conduction combination determination unit is used to determine the set of event-compatible conduction combinations supported by all states in each sub-state set of the certified state set, and to determine the intersection of the event-compatible conduction combination sets as the common event-compatible conduction combination set. Specifically, for each sub-state set in the certified state set, the pre-stored legal conduction combinations are traversed. For any legal conduction combination, it is determined whether all consecutive states in the sub-state set meet the phase sequence, commutation direction, rotor phase, and current flow direction requirements corresponding to the current machine-side line voltage zero-crossing event. A legal conduction combination is only included in the event-compatible conduction combination set corresponding to the sub-state set if all states in the sub-state set support the conduction combination. If a conduction combination is only applicable to some states in the sub-state set, it is not included in the corresponding event-compatible conduction combination set. After completing the determination of each sub-state set, the intersection of all event-compatible conduction combination sets is found. Each conduction combination in the resulting common event-compatible conduction combination set is supported by all possible states in the certified state set.

[0082] The identifier generation unit is used to update the current state set with the certified state set when both the certified state set and the common event compatible conduction combination set are not empty. This generates a one-time valid zero-crossing identifier that binds the machine-side line voltage zero-crossing event, the updated current state set, the common event compatible conduction combination set, the identifier valid time window, and the identifier usage status. Specifically, when both the certified state set and the common event compatible conduction combination set are not empty, the identifier generation unit replaces the current state set before event certification with the certified state set, ensuring that subsequent scheme generation is based on the state range verified by this zero-crossing event. The one-time valid zero-crossing identifier records at least the line voltage channel, zero-crossing direction, and timestamp of this machine-side line voltage zero-crossing event, and binds it to the updated current state set, the common event compatible conduction combination set, the identifier valid time window, and the identifier usage status. The starting point of the identifier valid time window is the time the one-time valid zero-crossing identifier is generated, and the ending point is no later than the earliest possible time of the next preset zero-crossing event, nor later than the termination time that the current commutation conditions can maintain. The identifier usage status is initially set to unused.

[0083] Within the range of conduction combinations allowed by a single valid zero-crossing identifier, the scheme generation module filters candidate conduction combinations based on the current state set, the turn-off and recovery state of the inverter bridge thyristors on the machine side, the current range of the DC reactor, and the trigger timing constraints. It then configures the target current path, common execution time window, and coordinated actions of rectification, inversion, and excitation for each candidate conduction combination to form a candidate commutation control scheme.

[0084] In one embodiment, the scheme generation module includes a conflict relationship construction unit, a combination filtering unit, and a scheme arrangement unit.

[0085] The conflict relationship construction unit is used to, for each sub-state set in the current state set, use each machine-side inverter bridge thyristor as a node, set the relationship between machine-side inverter bridge thyristor nodes in the same bridge arm as a fixed conflict relationship, and set a dynamic conduction conflict relationship based on the corresponding feasible machine-side inverter bridge thyristor conduction topology, machine-side inverter bridge thyristor turn-off recovery state, DC reactor current variation range, and time interval between trigger pulses of adjacent machine-side inverter bridge thyristors, thus constructing a corresponding dynamic conduction conflict relationship diagram. For example, a dynamic conduction conflict relationship diagram is constructed for each sub-state set in the current state set, where the six nodes correspond to... , , , , and . and , and , and Fixed conflict relationships are set between each bridge arm to prevent the upper and lower thyristors of the same bridge arm from conducting simultaneously. Dynamic conduction conflict relationships are determined based on the turn-off and recovery state of the inverter bridge thyristors on the machine side, the range of DC reactor current changes, the time interval between adjacent trigger pulses, and the candidate trigger times.

[0086] Let the turn-off time of a certain inverter bridge thyristor on the machine side be... The specified shutdown and recovery time is The safety margin for setting the time limit takes into account device variability, temperature variations, and control delays. When the candidate trigger time is... The difference is less than At that time, a dynamic conduction conflict relationship is set that relates to the thyristor being subjected to forward voltage or being re-conducted in advance.

[0087] When the time interval between the trigger pulses of the thyristors in adjacent inverter bridges is less than the preset minimum pulse interval, a dynamic conduction conflict relationship is set between the corresponding nodes. When the DC reactor current range at the candidate trigger time does not meet the allowable current range corresponding to the current commutation mode, a dynamic conduction conflict relationship is also set between the node pairs involved in the corresponding candidate conduction combination. Each dynamic conduction conflict relationship is associated with the corresponding sub-state set, the candidate trigger time interval, and the conflict establishment condition. The dynamic conduction conflict relationship is only valid within the corresponding condition and time interval. After the current state set, candidate trigger time, or control cycle changes, the dynamic conduction conflict relationship diagram is recalculated. The dynamic conduction conflict relationship diagram can be stored using a Boolean adjacency matrix. A matrix element of 1 indicates that the corresponding node pair has a conflict under the current sub-state set and candidate trigger time; a matrix element of 0 indicates that no corresponding conflict has been detected.

[0088] The combination filtering unit reads each conduction combination in the common event compatible conduction combination set one by one, and queries the dynamic conduction conflict relationship graph corresponding to each sub-state set in the current state set according to the triggering time to be used by the conduction combination. If any dynamic conduction conflict relationship graph contains a pair of conflicting nodes for the conduction combination, the conduction combination is deleted. Only conduction combinations that do not have any pairs of conflicting nodes in the dynamic conduction conflict relationship graphs corresponding to all sub-state sets are determined as candidate conduction combinations.

[0089] The scheme orchestration unit determines the time ranges for the thyristor turn-off recovery condition, current transfer condition, and excitation support condition to be met for each candidate conduction combination and each sub-state set. The intersection of these three time ranges for the same sub-state set yields the executable time range for that sub-state set; then, the intersection of the executable time ranges for all sub-state sets in the current state set yields a common execution time window. If the common execution time window is empty, the corresponding candidate conduction combination is deleted. Candidate conduction combinations correspond to the target current path. For example, candidate conduction combinations... The corresponding target current path is: the positive terminal of the DC reactor via... It enters the A-phase stator winding, and then passes through the C-phase stator winding. Return to the negative terminal of the DC reactor.

[0090] During the low-speed pulse commutation phase, the coordinated actions of rectification, inversion, and excitation can be arranged in the following relative order: adjust the control angle of the grid-side rectifier bridge to reduce the DC reactor current to the allowable commutation range; maintain or adjust the excitation current to ensure the terminal voltage meets the commutation support conditions; wait for the thyristors of the inverter bridge to be turned off to complete their turn-off recovery; output trigger pulses to the thyristors corresponding to the candidate conduction combinations; maintain the corresponding rectification and excitation actions within the preset confirmation time window until the target current path is established. After entering the load commutation phase, the back electromotive force of the synchronous motor can be used to complete the current transfer, while continuing to meet the turn-off recovery conditions and current transfer conditions of the inverter bridge thyristors on the machine side. The preset confirmation time window starts from the moment the trigger pulse of the target inverter bridge thyristor is output, and is determined based on the predicted current transfer time, thyristor turn-off recovery time, sampling delay, and safety margin. The candidate conduction combinations, the coordinated actions of rectification, inversion, and excitation within the joint execution time window, the target current path, and the preset confirmation time window together form the candidate commutation control scheme.

[0091] The safety verification module, based on a timing hybrid state model, predicts the state evolution of the current state set under various candidate commutation control schemes, forming a sequence of reachable state intervals. Commutation permission information is generated only when all reachable state intervals at each prediction time are within the safe state domain, and the final reachable state interval is within the target current path state domain. The safety verification module specifically includes a branch construction unit, a bidirectional reachability calculation unit, and a permission generation unit.

[0092] Among them, the branch construction unit is used to generate state evolution branches for each candidate commutation control scheme and each sub-state set in the current state set, based on the timing hybrid state model, according to the discrete mode conversion sequence within the common execution time window and the turn-off recovery time interval of the machine-side inverter bridge thyristor.

[0093] Specifically, the timing hybrid state model includes both continuous states and discrete modes. Continuous states include at least the DC reactor current, three-phase current, excitation current, rotor phase reference, rotor angular velocity, and the turn-off / recovery timing of the machine-side inverter bridge thyristors. Discrete modes include current conduction, rectification current reduction, machine-side inverter bridge thyristor turn-off / recovery, current transfer, target conduction, and commutation anomaly modes. Each state evolution branch is bound to a corresponding sub-state set, a discrete mode transition sequence, and the machine-side inverter bridge thyristor turn-off / recovery time interval.

[0094] The DC reactor current in each discrete mode is described by the following formula: ; For DC reactor inductance; This refers to the current in the DC reactor. For time; The rate of change of the DC reactor current with respect to time; For the grid-side rectifier bridge at the control angle The equivalent DC voltage generated below, where the subscript R indicates the grid-side rectifier bridge; For the machine-side inverter bridge, the thyristor conduction topology is as follows: and rotor phase reference The equivalent DC-side voltage is given by the subscript I, which indicates the inverter bridge on the machine side. This is the equivalent resistance of the DC circuit.

[0095] The three-phase currents on the generator side are represented by topology-dependent state equations: ; This is the three-phase stator inductance matrix for a synchronous motor. Let be the column vector of currents on the three-camera side, and The T in the upper right corner indicates the matrix transpose; The rate of change of the column vector of currents on the three-camera side with respect to time; For the topology turned on by the thyristors of the inverter bridge on the machine side A defined mapping matrix or mapping column vector from the DC-side voltage to the three-phase windings; This refers to the DC terminal voltage of the inverter bridge on the machine side. It refers to the stator resistance of a synchronous motor or a diagonal matrix composed of the stator resistances of each phase; The three-way back electromotive force column vector, For rotor phase reference, The rotor's electric angular velocity, This is the excitation current.

[0096] Different inverter bridge thyristor conduction topologies on the machine side correspond to different For example, by Towards During commutation, in the current transfer mode, both the original current branch of phase B and the target current branch of phase C are retained until the current of phase B is lower than the leakage current threshold. The remaining commutation process can establish corresponding topological mapping relationships according to the phase sequence.

[0097] The current conduction mode switches to rectification and current reduction mode after entering the common execution time window and starting rectification and current reduction actions; after the DC reactor current and the current of the branch to be turned off enter the preset turn-off range, it switches to the turn-off recovery mode of the inverter bridge thyristors on the machine side; after the turn-off recovery time meets the device requirements and the target trigger pulse has been output, it switches to current transfer mode; after the target current path is established, it switches to target conduction mode. In case of bridge arm shoot-through, overcurrent, insufficient turn-off recovery, incorrect current transfer direction, or confirmation timeout, it switches to commutation abnormality mode.

[0098] For each candidate commutation control scheme and each sub-state set, the branching unit expands the state evolution branch according to the discrete mode transition sequence that may occur within the common execution time window and the turn-off recovery time interval of the machine-side inverter bridge thyristors. When a certain discrete mode transition condition may be met within a time interval, the time interval is divided into multiple sub-intervals according to a preset time resolution, and each sub-interval forms a state evolution branch. Each state evolution branch is bound to the corresponding candidate commutation control scheme, the initial sub-state set, the corresponding feasible machine-side inverter bridge thyristor conduction topology, the discrete mode transition sequence, and the turn-off recovery time interval of the machine-side inverter bridge thyristors.

[0099] The bidirectional reachability calculation unit is used to calculate the reachable state interval sequence in the forward direction along each state evolution branch, and to calculate the reverse reachable state interval of each state evolution branch at each prediction time in the reverse direction from the target current path state domain according to the remaining rectification, inversion and excitation coordinated actions of the corresponding candidate commutation control scheme after each prediction time.

[0100] During forward computation, a preset confirmation time window is divided into multiple prediction steps, and the set of sub-states bound to each state evolution branch is used as the initial reachable state interval. This implementation uses interval Taylor integration to calculate the forward reachable state interval. First, the initial sub-state set is represented as an axis-aligned interval box, and the synchronous motor parameters, DC circuit parameters, thyristor device parameters, and unmodeled disturbances are represented as intervals. Then, within each prediction step, the state change range is calculated according to the state equation corresponding to the current discrete mode, and the integral remainder, rounding error, and disturbance range are added to the prediction result. After each prediction step, the discrete mode transition conditions are checked. When multiple transition conditions may be met, the current reachable state interval is copied and different state evolution branches are formed. When the interval width of a certain state dimension exceeds the preset width, the state dimension is bisected to control the excessive expansion of the interval calculation.

[0101] The reverse calculation uses the target current path state domain as the final state range. For the state transition relationship between adjacent prediction times, the state at the next prediction time is limited to the determined reverse reachable state interval. The coordinated actions of rectification, inversion, and excitation are fixed to the action range specified by the corresponding candidate commutation control scheme in this prediction step, and the equipment parameters are limited to the corresponding parameter intervals. The intersection of the safe state domain and the forward reachable state interval of the previous prediction time is used as the initial search range. The state range of the previous prediction time is shrunk using the interval Newton method or the forward-reverse interval shrinkage method. The interval retained after shrinkage is determined as the reverse reachable state interval of the corresponding prediction time. When the shrinkage result is empty, it means that the state evolution branch continues to execute the original candidate commutation control scheme from the corresponding prediction time and can no longer enter the target current path state domain. During the reverse calculation process, the candidate commutation control scheme, the type of coordinated action, and the order of coordinated actions remain unchanged.

[0102] The permit generation unit is used to determine the commutation permit information generation conditions as follows: the reachable state intervals of each state evolution branch at each prediction time are all located within the intersection of the safe state domain and the corresponding reverse reachable state interval, and the terminal reachable state intervals are all located within the target current path state domain. When the commutation permit information generation conditions are met, commutation permit information is generated that binds the corresponding candidate commutation control scheme, the one-time effective zero-crossing identifier, and the set of state evolution branches.

[0103] Specifically, the safety state domain is established based on the hard safety limits of the inverter bridge thyristors, synchronous motor, and DC circuit on the machine side, excluding at least the states where thyristors on the same bridge arm are simultaneously turned on, DC reactor current overcurrent, machine-side phase current overcurrent, insufficient thyristor turn-off recovery time, incorrect current transfer direction, and entering the abnormal commutation mode. The target current path state domain is established based on the corresponding candidate commutation control scheme, limiting at least the inflow phase current range, outflow phase current range, upper limit of non-target phase leakage current, DC reactor current range, and rotor phase range in the target current path. To ensure that the permissible judgment covers all state evolution branches, all prediction moments, and all safety limits, this implementation method judges the first state according to the following formula. Do the candidate commutation control schemes meet the conditions for generating commutation permission information?

[0104] ;

[0105] In the formula, For the first The approval decision results for each candidate commutation control scheme. express Each candidate commutation control scheme meets the conditions for generating commutation permission information; For the first The set of state evolution branches corresponding to each candidate commutation control scheme. Number the branches of the state evolution; For the prediction step number, This refers to the total number of prediction steps or the sequence number of the final prediction step corresponding to the preset confirmation time window. For the first The first candidate commutation control scheme The state evolution branch is at the _th ... The positive reachable state interval at each predicted time. Represents the set of forward reachable states; For the same candidate commutation control scheme, the same state evolution branch, and the same prediction time, the reverse reachable state interval is defined. Represents the set of states that can be reached in the reverse direction; A security state domain constructed according to hard security constraints; For the first The target current path state domain corresponding to each candidate commutation control scheme; It can be any state within the positive reachable state range; Security restriction number; The total number is limited for safety purposes; For state Compared to the first The signed safety margin of the safety limit boundary is positive when the state is on the permissive side; For the first The minimum safety margin required to be maintained under each safety restriction; For the first The normalized scale corresponding to the safety margin, and with and They have the same dimensions; inf represents taking the infimum within the positive reachable state interval; min represents taking the minimum value among all state evolution branches, all prediction steps, and all safety constraints; This indicates that the corresponding condition applies to all subsequent objects. Indicates a set containment relationship; Represents the intersection of sets; This indicates a logical AND operation.

[0106] The first requirement is that the reachable state intervals of all state evolution branches at all predicted times lie at the intersection of the safe state domain and the corresponding reverse reachable state interval. The second requirement is that the final reachable state intervals of all state evolution branches lie within the corresponding target current path state domain. The third requirement is that all reachable states retain a normalized margin no less than the corresponding minimum safety margin relative to each safety limit. When all conditions are met, the permission generation unit generates commutation permission information and binds this information to the corresponding candidate commutation control scheme, the one-time valid zero-crossing identifier, and the set of verified state evolution branches.

[0107] Before execution, the execution and rollback modules re-verify whether the actual operating state is still consistent with the state conditions used when the commutation permission information was generated. After verification, the permitted candidate commutation control scheme is executed. During execution, the actual state trajectory is continuously reconstructed. If the actual state trajectory exceeds the predicted reachable state range, or if the target current path state domain is not entered by the end of the preset confirmation time window, a safety rollback operation is performed. The safety rollback operation includes locking the thyristor trigger pulse of the inverter bridge on the machine side, adjusting the control angle of the grid-side rectifier bridge to reduce the DC reactor current, and restoring the control state machine to the commutation control state saved when the commutation was last confirmed successfully. The commutation control state includes at least the commutation sequence number, confirmed conduction combination, operating direction, commutation mode, and corresponding rectification and excitation control references.

[0108] Furthermore, the commutation authorization information also includes the authorization validity period window and authorization status, which includes at least an unused status, a used status, and an invalid status. The authorization validity period window is jointly defined by the authorization validity period window of a one-time valid zero-crossing identifier, the common execution period window of candidate commutation control schemes, and the expected retention time of the status conditions used when generating the commutation authorization information.

[0109] The execution and rollback module includes a license verification unit. Before execution, the license verification unit calls the state construction module to generate a real-time state set based on the sampled data at the execution entry point. The real-time state set uses the same state estimation, conduction topology verification, and state set generation process as the current state set, but it is used to characterize the latest operating state after the commutation license information is generated, and does not cover the current state set on which the commutation license information was based. The real-time machine-side inverter bridge thyristor conduction topology is determined according to the aforementioned conduction topology verification method, based on the machine-side inverter bridge thyristor conduction state, trigger pulse readback, and machine-side phase current direction collected before execution.

[0110] The permission verification unit reads the state evolution branches bound to the commutation permission information one by one. For each state evolution branch, it determines whether the real-time state set is completely contained within the sub-state set bound to that state evolution branch, and whether the real-time machine-side inverter bridge thyristor conduction topology is consistent with the feasible machine-side inverter bridge thyristor conduction topology corresponding to that sub-state set. The state set inclusion relationship and the conduction topology consistency relationship must be satisfied simultaneously by the same state evolution branch. If one state evolution branch satisfies the state set inclusion relationship and another state evolution branch satisfies the conduction topology consistency relationship, then the execution entry verification condition is not met. After the existence of the same state evolution branch, it continues to verify whether the current time is within the valid time window of the one-time valid zero-crossing flag, whether the flag usage state is unused, whether the current time is within the valid permission time window, and whether the permission state is unused. When all conditions are met, an execution permission signal is generated for the corresponding candidate commutation control scheme.

[0111] The execution and rollback module also includes a scheme optimization unit, which is used to normalize the distance from the reachable state interval to the boundary of the safe state domain and the distance to the boundary of the corresponding reverse reachable state interval of each state evolution branch at each prediction time according to the preset state scale corresponding to each state variable for each candidate commutation control scheme that has generated execution permission signal. The minimum value of the two types of normalized distances in all state evolution branches and all prediction times is determined as the scheme robustness margin, and the candidate commutation control scheme with the largest scheme robustness margin is determined as the target commutation control scheme.

[0112] Specifically, each state variable has different dimensions. The scheme optimization unit sets preset state scales for DC reactor current, machine-side phase current, excitation current, rotor phase reference, rotor angular velocity, and turn-off recovery time. The preset state scales adopt the rated value, allowable variation range, or safety limit range of the corresponding state variable and have the same dimensions as the corresponding state variable. Each boundary of the safe state domain is represented by a signed distance function. The signed distance is positive when the reachable state interval is completely within the safe state domain; it is zero when it touches the boundary of the safe state domain. Dividing the signed distance by the corresponding preset state scale yields the normalized distance to the boundary of the safe state domain. When the reachable state interval and the reverse reachable state interval are represented using axis-aligned interval boxes, the normalized distance to the boundary of the reverse reachable state interval is determined based on the minimum distance from the upper and lower boundaries of each state dimension of the reachable state interval to the corresponding upper and lower boundaries of the reverse reachable state interval, and then divided by the preset state scale of the corresponding state variable.

[0113] No. The robustness margin of each candidate commutation control scheme is determined by the following formula: ; For the first The robustness margin of each candidate commutation control scheme is a dimensionless quantity. For the first The first candidate commutation control scheme The state evolution branch is at the _th ... The normalized distance from the prediction time to the boundary of the safe state domain; It is the normalized distance from the boundary of the same reachable state interval to the corresponding reverse reachable state interval; This indicates that the minimum value is taken among all state evolution branches and all prediction times of the corresponding candidate commutation control scheme.

[0114] The scheme selection unit compares the scheme robustness margins of each candidate commutation control scheme that has generated an execution permission signal, and determines the candidate commutation control scheme with the largest scheme robustness margin as the target commutation control scheme. The commutation permission information generation conditions are used to determine whether the candidate commutation control scheme meets the minimum safety requirements; the scheme robustness margin is used to select the scheme with the larger safety margin and target achievable margin among the candidate commutation control schemes that have met the minimum safety requirements.

[0115] Within the shared execution time window of the target commutation control scheme, the execution and rollback modules control the rectification, inversion, and excitation actions according to the types and sequences specified in the target commutation control scheme. After outputting the trigger pulse to the thyristor of the inverter bridge on the target unit side, the module updates the permission status of the commutation permission information corresponding to the target commutation control scheme from unused to used, and updates the usage status of the one-time valid zero-crossing flag bound to the commutation permission information from unused to used. The permission status and flag usage status are updated using an indivisible status update operation to prevent the same commutation permission information or one-time valid zero-crossing flag from being reused.

[0116] The execution and rollback module also includes a trajectory reconstruction unit, a branch monitoring unit, a path confirmation unit, a rollback control unit, an error boundary correction unit, and a scheme isolation unit.

[0117] After the target commutation control scheme is executed, the trajectory reconstruction unit calls the state construction module at each sampling moment to generate the real-time state set for that sampling moment. The real-time machine-side inverter bridge thyristor conduction topology is determined according to the aforementioned conduction topology verification method, based on the machine-side inverter bridge thyristor conduction state, trigger pulse readback, and machine-side phase current direction at the corresponding sampling moment. The real-time state set, real-time machine-side inverter bridge thyristor conduction topology, and timestamps for each sampling moment are saved in chronological order to form the actual state trajectory. When the sampling moment coincides with the predicted moment in the forward reachable state interval sequence, the reachable state interval of the corresponding predicted moment is directly used. When the sampling moment is between two adjacent predicted moments, the interval envelope formed by the two adjacent reachable state intervals and the time alignment error is used as the reachable state interval corresponding to that sampling moment.

[0118] At the first sampling moment after the execution of the target commutation control scheme, the branch monitoring unit determines all state evolution branches bound to the target commutation control scheme as the current retention set. At each sampling moment, only state evolution branches whose machine-side inverter bridge thyristor conduction topology is consistent with the real-time machine-side inverter bridge thyristor conduction topology are selected from the current retention set to form a topology-consistent state evolution branch set. State evolution branches that have been excluded at previous sampling moments will not participate in subsequent selection. The branch monitoring unit reads the reachable state intervals of each topology-consistent state evolution branch at the corresponding prediction moment and calculates the union of these reachable state intervals to form a topology-consistent reachable state interval union. Then, it determines whether the real-time state set is completely contained within the topology-consistent reachable state interval union. Complete containment means that all possible states in the real-time state set can be explained by at least one topology-consistent state evolution branch. When complete coverage is achieved, state evolution branches whose reachable state intervals intersect with the real-time state set are retained from the topology-consistent state evolution branch set, and these state evolution branches are determined as the current retention set for the next sampling moment. Therefore, the current retained set gradually shrinks with the actual state trajectory, and the state evolution branches that are no longer compatible with the historical actual state trajectory cannot re-enter the current retained set at subsequent sampling times.

[0119] When reachable state intervals are represented using interval boxes, a recursive interval partitioning method is used to determine whether the real-time state set is completely covered by the union of topologically consistent reachable state intervals. First, the real-time state set is divided into multiple sub-interval boxes. If a sub-interval box is completely contained within at least one reachable state interval, it is considered covered. If a sub-interval box does not intersect with any of the reachable state intervals, it is considered uncovered. In other cases, the sub-interval box is further bisected. If, after subdividing to a preset minimum resolution, a sub-interval box still cannot be confirmed as completely covered, it is treated as uncovered to avoid incorrect coverage determinations due to numerical errors.

[0120] The path confirmation unit is used to determine whether the actual state trajectory has entered the target current path state domain within a preset confirmation time window, provided that the real-time state set is included in the target current path state domain and the real-time machine-side inverter bridge thyristor conduction topology is consistent with the machine-side inverter bridge thyristor conduction topology corresponding to the candidate conduction combination. Specifically, the target current path state domain is used to define the continuous electrical state after commutation, including the target inflow phase current range, the target outflow phase current range, the upper limit of non-target phase leakage current, the DC reactor current range, and the rotor phase range.

[0121] With candidate conduction combination For example, the target current path requires that the current on camera A is within the positive target current range, the current on camera C is within the negative target current range, the current on camera B is below the upper limit of the non-target phase leakage current, and the DC reactor current and rotor phase reference are within the corresponding allowable range.

[0122] Within a preset confirmation time window, the path confirmation unit simultaneously checks whether the real-time state set is completely contained within the target current path state domain, and whether the real-time machine-side inverter bridge thyristor conduction topology is consistent with the machine-side inverter bridge thyristor conduction topology corresponding to the candidate conduction combinations in the target commutation control scheme. When both conditions are met, the actual state trajectory is determined to have entered the target current path state domain.

[0123] The rollback control unit is used to perform a safe rollback operation when the set of evolutionary branches of the topology-consistent state is empty, the real-time state set is not included in the union of the topology-consistent reachable state intervals, or the confirmation conditions of the path confirmation unit are not met.

[0124] An empty set of topology-consistent state evolution branches indicates that the actual inverter bridge thyristor conduction topology cannot be explained by any currently retained state evolution branch; a real-time state set not included in the union of topology-consistent reachable state intervals indicates that the actual continuous state exceeds the state range predicted during permissible generation; and failure to meet the path confirmation condition by the end of the preset confirmation time window indicates that the target current path has not been established as expected. In any of these cases, the thyristor trigger pulses of the inverter bridge on the machine side are first blocked, and then the control angle of the grid-side rectifier bridge is adjusted to reduce the DC reactor current according to the preset decreasing slope. After the DC reactor current falls below the safety threshold, the commutation sequence number, conduction combination record, and control state machine are restored to the commutation control state saved at the time of the most recent successful commutation confirmation.

[0125] Error boundary correction is initiated only when the set of evolutionary branches of the topologically consistent state is not empty, but the real-time state set is not completely covered by the union of the topologically consistent reachable state intervals. This condition indicates that the actual thyristor conduction topology of the actual machine-side inverter bridge can still be explained by the predicted branches, and the insufficient coverage mainly occurs in the continuous state range. The error boundary correction unit searches for the first real-time state set with insufficient coverage according to the sampling time sequence, and compares the projection of this real-time state set with the union of the topologically consistent reachable state intervals on each state dimension. When the upper bound of the real-time state set on a certain state dimension exceeds the upper bound of the union of the topologically consistent reachable state intervals, the deviation direction is determined as the upper bound direction; when the lower bound of the real-time state set on a certain state dimension is lower than the lower bound of the union of the topologically consistent reachable state intervals, the deviation direction is determined as the lower bound direction. When multiple state quantities have insufficient coverage at the same time, the ratio of the excess amount of each state quantity to the corresponding preset state scale is calculated, and the state quantity with the largest ratio is selected for correction first. After determining the state quantity to be corrected and the deviation direction, the error boundaries of other state quantities remain unchanged, and only the error boundaries of the selected state quantity on the corresponding deviation direction are expanded. The minimum boundary correction is the minimum increase required to ensure that the recalculated union of topologically consistent reachable state intervals completely covers all existing real-time state sets. The minimum boundary correction can be determined using a binary search. During the search, the discrete mode transition sequence of the original topologically consistent state evolution branches remains unchanged; only the initial state interval corresponding to the selected state variable is expanded, and an interval reachability calculation method satisfying set inclusion relations is used. Under this condition, the recalculated reachable state envelope does not shrink when the error boundary correction increases. When new discrete mode transition possibilities arise after the error boundary is expanded, a corresponding new state evolution branch is generated. The reachable state interval of the new state evolution branch is added to the union of topologically consistent reachable state intervals, and the minimum boundary correction search restarts. If complete coverage cannot be achieved after reaching the preset maximum allowable correction, the expansion of the state variable error boundary stops. The error boundary correction unit sends the state variable to be corrected, the deviation direction, and the minimum boundary correction to the state construction module. The state construction module corrects the corresponding state variable error boundaries, regenerates each sub-state set, and updates the current state set.

[0126] The license verification unit checks for phase-change license information that has not yet been executed. When the set of state evolution branches bound to the phase-change license information is generated based on the current state set before the update, the license status of the phase-change license information is updated to invalid to prevent the continued use of phase-change license information generated based on the old state boundary.

[0127] The scheme isolation unit is used to generate scheme feature identifiers based on the candidate conduction combinations, target current paths, and the type and order of rectification, inversion, and excitation coordinated actions in the target commutation control scheme. The safety verification module performs non-executable re-verification of candidate commutation control schemes with scheme feature identifiers based on the updated current state set. If a candidate commutation control scheme still does not meet the conditions for generating commutation permission information, the scheme feature identifier is added to the scheme isolation set, prohibiting the candidate commutation control scheme with scheme feature identifiers from participating in the subsequent generation of commutation permission information and determination of the target commutation control scheme.

[0128] Specifically, the scheme feature identifier consists of structural features that can be repeated during subsequent commutation processes, including candidate conduction combinations, target current paths, rectification action types, inverter action types, excitation action types, and the relative order of these actions. The common execution time window and absolute trigger moment are related to a single zero-crossing event and are not part of the scheme feature identifier. For example, a scheme feature can be represented as: performing rectification and current reduction, waiting for the machine-side inverter bridge thyristors to complete turn-off recovery, and triggering... Maintain excitation, with the target current path being A-phase inflow and C-phase outflow.

[0129] After the current state set is updated, the safety verification module performs non-executable re-verification of candidate commutation control schemes with corresponding scheme feature identifiers, following the state evolution branch construction process, the forward reachable state interval calculation process, the reverse reachable state interval calculation process, and the commutation permission information generation conditions. Non-executable re-verification only performs calculations and judgments; it does not output machine-side inverter bridge thyristor trigger pulses, does not generate execution permission signals for actual execution, and does not change the usage status of the one-time valid zero-crossing identifier. If a candidate commutation control scheme with corresponding scheme feature identifiers still does not meet the commutation permission information generation conditions, its scheme feature identifier is added to the scheme isolation set. Subsequently, the scheme isolation set is queried before generating commutation permission information and before the scheme optimization unit determines the target commutation control scheme. Candidate commutation control schemes with isolated scheme feature identifiers do not participate in subsequent commutation permission information generation or target commutation control scheme determination.

[0130] like Figure 2 As shown, another embodiment of the present invention provides a commutation control method for a static start frequency converter, which is executed cyclically by a controller according to a control cycle, including the following steps:

[0131] S1: Collects machine-side line voltage, machine-side phase current, DC reactor current, excitation current, and the conduction status of the machine-side inverter bridge thyristors. Performs time synchronization, zero-bias correction, and filtering on each signal. Determines the rotor phase reference based on the machine-side line voltage and machine-side phase current, and sets the state quantity error boundary in combination with measurement error, filtering delay, and equipment parameter error to form the current possible state range, i.e., the current state set. The current state set includes both the continuous state quantity range and the conduction status of the machine-side inverter bridge thyristors compatible with the collected data.

[0132] S2: Detect zero-crossing events of the machine side line voltage, record the zero-crossing channel, zero-crossing direction, and zero-crossing time; determine whether the current zero-crossing event is the one that should occur based on the preset phase sequence; determine whether the angle relationship is consistent based on the zero-crossing time and rotor phase reference; when the phase sequence, time, and angle consistency are met, generate a one-time valid zero-crossing identifier, and set the identifier's valid time window and unused state; if any authentication condition is not met, discard the zero-crossing event and do not proceed to the commutation scheme generation step.

[0133] S3: Based on the commutation direction defined by the one-time valid zero-crossing identifier and the current state set, determine the range of event-compatible conduction combinations. Construct a dynamic conduction conflict relationship diagram based on the turn-off recovery state of the thyristors of the machine-side inverter bridge, the current range of the DC reactor current, and the trigger timing constraints, and delete the conduction combinations with conflicts. For the remaining candidate conduction combinations, determine the target current path, and arrange the coordinated actions of the grid-side rectifier bridge, the machine-side inverter bridge, and the excitation device to form a candidate commutation control scheme.

[0134] S4: Establish a timing hybrid state model describing the changes in DC reactor current, machine-side phase current, excitation current, rotor phase, and machine-side inverter bridge conduction state. For each candidate commutation control scheme, calculate the sequence of reachable state intervals within a preset confirmation time window starting from the current state set. Determine whether each reachable state interval in the sequence is located in the safe state domain, and whether the final reachable state interval is located in the target current path state domain. If all conditions are met, generate commutation permission information that binds the corresponding candidate commutation control scheme and a one-time valid zero-crossing identifier; otherwise, delete the candidate commutation control scheme.

[0135] S5: Before execution, check again whether the one-time valid zero-crossing indicator and commutation permission information are both in a valid state and an unused state. After verification, execute the rectification, inversion and excitation coordinated action according to the candidate commutation control scheme bound to the commutation permission information; after the thyristor trigger pulse output of the inverter bridge on the machine side, update the one-time valid zero-crossing indicator and commutation permission information to the used state.

[0136] During execution, the actual state trajectory is continuously constructed and compared with the reachable state interval sequence corresponding to the commutation permission information. If the actual state trajectory exceeds the reachable state interval sequence, or if it does not enter the target current path state domain at the end of the preset confirmation time window, the trigger pulse of the inverter bridge thyristor on the machine side is immediately blocked, the DC reactor current is controlled to decrease to the safety threshold, and the control state machine is rolled back to the commutation control state when the commutation was most recently confirmed to be successful.

[0137] The above steps are repeated each time a commutation event occurs. The one-time valid zero-crossing identifier and commutation permission information only apply to the current commutation process they are bound to and cannot be reused across zero-crossing events.

[0138] In summary, this invention constrains the entire commutation process, which is subject to measurement errors, parameter deviations, and uncertainties in conduction topology, through state set construction, zero-crossing event authentication, candidate commutation control scheme screening, reachable state safety verification, and execution process backoff control. This reduces zero-crossing misjudgments, thyristor false triggering, commutation failures, and current surges, thereby improving the safety, robustness, and continuous operation capability of the static start-up inverter during startup.

[0139] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A commutation control system for a static start frequency converter, characterized in that, include: The state construction module is used to construct a current state set, including rotor phase reference and defined by state quantity error boundaries, based on the machine-side line voltage, machine-side phase current, DC reactor current, excitation current and the conduction state of the machine-side inverter bridge thyristors. The event authentication module is used to generate a one-time valid zero-crossing identifier based on the phase sequence, time, and angular consistency with the rotor phase reference of the zero-crossing event of the machine-side line voltage. The scheme generation module is used to screen candidate conduction combinations based on the one-time effective zero-crossing identifier, the current state set, and the turn-off recovery state of the inverter bridge thyristors on the machine side, using a dynamic conduction conflict relationship diagram, and to combine the candidate conduction combinations, target current paths, and rectifier, inverter, and excitation coordinated actions into candidate commutation control schemes. The safety verification module is used to calculate the sequence of reachable state intervals of the current state set under each candidate commutation control scheme based on the timing hybrid state model. When each reachable state interval in the sequence is located in the safe state domain and the last reachable state interval is located in the target current path state domain, it generates a binding one-time valid zero-crossing identifier and commutation permission information of the candidate commutation control scheme. The execution and rollback module is used to execute the bound candidate commutation control scheme when both the one-time valid zero-crossing identifier and commutation permission information are valid and unused; When the actual state trajectory exceeds the reachable state interval sequence or fails to enter the target current path state domain within the preset confirmation time window, a safety backoff operation is performed, including locking the inverter bridge thyristor trigger pulse on the machine side, reducing the DC reactor current to a safe threshold, and backoffing to the commutation control state when the most recent commutation was successful.

2. The commutation control system for a static start frequency converter according to claim 1, characterized in that, The state construction module includes: The state estimation unit is used to determine the estimated values ​​of basic state variables, including rotor phase reference and rotor angular velocity, based on the machine-side line voltage, machine-side phase current, DC reactor current and excitation current. The conduction topology verification unit is used to determine the expected conduction topology of the machine-side inverter bridge thyristors based on the triggering records of the machine-side inverter bridge thyristors, determine the observed current path based on the current continuity relationship between the machine-side phase current and the DC reactor current, and determine the set of feasible machine-side inverter bridge thyristor conduction topologies based on the compatibility between the expected conduction topology of the machine-side inverter bridge thyristors and the observed current path, as well as the unidirectional conduction constraints, bridge arm conduction constraints and current continuity constraints of the machine-side inverter bridge thyristors. The state set generation unit is used to determine the corresponding state quantity error boundary based on the basic state quantity estimate for each feasible machine-side inverter bridge thyristor conduction topology, and to constrain and correct the state range defined by the state quantity error boundary according to the current continuity relationship between the machine-side phase current and the DC reactor current, the phase relationship between the machine-side line voltage and the rotor phase reference, and the allowable conversion relationship of the machine-side inverter bridge thyristor conduction topology in adjacent control cycles, thereby generating a sub-state set corresponding to each feasible thyristor conduction topology, and determining the union of each sub-state set as the current state set.

3. The commutation control system for a static start frequency converter according to claim 2, characterized in that, The event authentication module includes: The event matching unit is used to generate a predicted zero-crossing window for each sub-state set in the current state set based on the corresponding feasible machine-side inverter bridge thyristor conduction topology, rotor phase reference, rotor angular velocity and preset zero-crossing phase sequence, and determine the event matching state set based on the matching results between the machine-side line voltage zero-crossing event and each predicted zero-crossing window. The topology verification unit is used to determine the event matching state set as the authentication state set; when there is a previously generated one-time valid zero-crossing identifier, it determines the set of thyristor conduction topologies that can reach the machine-side inverter bridge based on the current state set bound to the previously generated one-time valid zero-crossing identifier and the preset allowed commutation sequence, and filters the authentication state set accordingly. The common conduction combination determination unit is used to determine the event-compatible conduction combination set supported by each sub-state set in the authentication state set in all states, and to determine the intersection of each event-compatible conduction combination set as the common event-compatible conduction combination set. The identifier generation unit is used to update the current state set with the authentication state set when both the authentication state set and the common event compatible conduction combination set are not empty, and to generate a zero-crossing event of the bonding machine side line voltage, the updated current state set, the common event compatible conduction combination set, the identifier valid time window, and a one-time valid zero-crossing identifier of the identifier usage status.

4. The commutation control system for a static start frequency converter according to claim 3, characterized in that, The scheme generation module includes: The conflict relationship construction unit is used to set the relationship between the machine-side inverter bridge thyristor nodes in the same bridge arm as a fixed conflict relationship for each sub-state set in the current state set, with each machine-side inverter bridge thyristor as a node, and to set a dynamic conduction conflict relationship according to the corresponding feasible machine-side inverter bridge thyristor conduction topology, machine-side inverter bridge thyristor turn-off recovery state, DC reactor current variation range, and time interval between adjacent machine-side inverter bridge thyristor trigger pulses, and to construct a corresponding dynamic conduction conflict relationship diagram. The combination filtering unit is used to exclude, from the set of common event compatible conduction combinations bound by the one-time valid zero-crossing identifier, conduction combinations that contain conflicting node pairs in any dynamic conduction conflict relationship graph, and determine the remaining conduction combinations as the candidate conduction combinations; The scheme arrangement unit is used to determine, for each candidate conduction combination, the common execution time window in which each sub-state set in the current state set satisfies the turn-off recovery condition, current transfer condition, and excitation support condition of the inverter bridge on the machine side, and to determine the target current path corresponding to each candidate conduction combination. The candidate conduction combination, the rectification, inversion and excitation coordinated action within the common execution time window, the target current path and the preset confirmation time window are combined into the candidate commutation control scheme.

5. The commutation control system for a static start frequency converter according to claim 4, characterized in that, The security verification module includes: The branch construction unit is used to generate state evolution branches based on the timing hybrid state model, according to the discrete mode conversion sequence within the common execution time window and the turn-off recovery time interval of the machine-side inverter bridge thyristors, for each candidate commutation control scheme and each sub-state set in the current state set. The discrete modes of the timing hybrid state model include current conduction, rectification current reduction, machine-side inverter bridge thyristor turn-off recovery, current transfer, target conduction, and commutation anomaly mode. Each state evolution branch is bound to the corresponding sub-state set, discrete mode conversion sequence, and machine-side inverter bridge thyristor turn-off recovery time interval. The bidirectional reachability calculation unit is used to calculate the reachable state interval sequence in the forward direction along each state evolution branch, and to calculate the reverse reachable state interval of each state evolution branch at each prediction time in the reverse direction from the target current path state domain according to the remaining rectification, inversion and excitation coordinated action of the corresponding candidate commutation control scheme after each prediction time. The permit generation unit is used to determine the commutation permit information generation condition as follows: the reachable state intervals of each state evolution branch at each prediction time are all located within the intersection of the safe state domain and the corresponding reverse reachable state interval, and the terminal reachable state intervals are all located within the target current path state domain. When the commutation permit information generation condition is met, commutation permit information is generated that binds the corresponding candidate commutation control scheme, the one-time effective zero-crossing identifier, and the set of state evolution branches.

6. The commutation control system for a static start frequency converter according to claim 5, characterized in that, The phase-switching license information also includes the license validity period window and the license status; The execution and rollback module includes a permission verification unit, which is used to call the state construction module before execution to generate a real-time state set for execution entry verification, and determine the real-time machine-side inverter bridge thyristor conduction topology based on the thyristor conduction status of the machine-side inverter bridge collected before execution; determine whether there is a common state evolution branch from the state evolution branch set bound to the commutation permission information, so that the real-time state set is included in the sub-state set bound to the state evolution branch, and the real-time machine-side inverter bridge thyristor conduction topology is consistent with the feasible machine-side inverter bridge thyristor conduction topology corresponding to the sub-state set; when there is a common state evolution branch, the one-time valid zero-crossing identifier bound to the commutation permission information is within the identifier valid time window and the identifier usage status is unused, and the commutation permission information is within the permission valid time window and the permission status is unused, an execution permission signal is generated for the corresponding candidate commutation control scheme.

7. The commutation control system for a static start frequency converter according to claim 6, characterized in that, The execution and rollback module also includes a scheme optimization unit; The scheme selection unit is used to normalize the distance from the reachable state interval of each state evolution branch to the boundary of the safe state domain and the distance to the boundary of the corresponding reverse reachable state interval at each prediction time for each candidate commutation control scheme that has generated an execution permission signal, according to the preset state scale corresponding to each state variable. The minimum value of the two types of normalized distances in all state evolution branches and all prediction times is determined as the scheme robustness margin, and the candidate commutation control scheme with the largest scheme robustness margin is determined as the target commutation control scheme. The execution and rollback module executes within the common execution time window corresponding to the target commutation control scheme, and after the output side inverter bridge thyristor trigger pulse, updates the permission status and usage status corresponding to the target commutation control scheme to the used status respectively.

8. The commutation control system for a static start frequency converter according to claim 7, characterized in that, The execution and rollback module also includes: The trajectory reconstruction unit is used to call the state construction module at each sampling time to generate a real-time state set after the target commutation control scheme is executed, and to determine the real-time machine-side inverter bridge thyristor conduction topology based on the machine-side inverter bridge thyristor conduction state at each sampling time, thereby forming the actual state trajectory. The branch monitoring unit is used to determine the set of state evolution branches corresponding to the target commutation control scheme as the current retained set at the first sampling time. At each sampling time, it determines the set of topologically consistent state evolution branches that are consistent with the real-time thyristor conduction topology of the machine-side inverter bridge from the current retained set, and determines the union of the reachable state intervals of the set of topologically consistent state evolution branches at the corresponding prediction time as the union of topologically consistent reachable state intervals. When the real-time state set is included in the union of topologically consistent reachable state intervals, the topologically consistent state evolution branch that intersects with the reachable state interval and the real-time state set is determined as the current retained set at the next sampling time. The path confirmation unit is used to determine that the actual state trajectory enters the target current path state domain when the real-time state set is included in the target current path state domain and the real-time machine-side inverter bridge thyristor conduction topology is consistent with the machine-side inverter bridge thyristor conduction topology corresponding to the candidate conduction combination within a preset confirmation time window. The rollback control unit is used to execute the safe rollback operation when the set of evolutionary branches of the topology-consistent state is empty, the real-time state set is not included in the union of the topology-consistent reachable state intervals, or the confirmation conditions of the path confirmation unit are not met.

9. The commutation control system for a static start frequency converter according to claim 8, characterized in that, The execution and rollback module also includes: An error boundary correction unit is used to determine the state variables and deviation directions that were not fully covered initially after the rollback control unit performs a safe rollback operation due to the non-empty set of topologically consistent state evolution branches and the real-time state set not being included in the union of topologically consistent reachable state intervals. Under the condition that the error boundaries of other state variables remain unchanged, it determines the minimum boundary correction amount required to ensure that the recalculated union of topologically consistent reachable state intervals fully covers each formed real-time state set, and sends the minimum boundary correction amount to the state construction module. The state construction module corrects the corresponding state variable error boundaries according to the minimum boundary correction amount and updates the current state set. The permission verification unit updates the permission status of the phase commutation permission information generated based on the current state set before the update to an invalid status. The scheme isolation unit is used to generate scheme feature identifiers based on the candidate conduction combinations, target current paths, and the type and order of rectification, inversion, and excitation coordinated actions in the target commutation control scheme. The safety verification module performs non-executable re-verification of candidate commutation control schemes with scheme feature identifiers based on the updated current state set. If a candidate commutation control scheme still does not meet the conditions for generating commutation permission information, the scheme feature identifier is added to the scheme isolation set, prohibiting the candidate commutation control scheme with the scheme feature identifier from participating in the subsequent generation of commutation permission information and determination of the target commutation control scheme.

10. A commutation control method for a static start frequency converter, characterized in that, include: Based on the machine-side line voltage, machine-side phase current, DC reactor current, excitation current, and the conduction state of the machine-side inverter bridge thyristors, a current state set is constructed, including the rotor phase reference and defined by the state quantity error boundary. A one-time valid zero-crossing identifier is generated based on the phase sequence, time, and angle consistency with the rotor phase reference of the zero-crossing event of the machine-side line voltage. Based on the one-time effective zero-crossing identifier, the current state set, and the turn-off recovery state of the inverter bridge thyristors on the machine side, candidate conduction combinations are screened using a dynamic conduction conflict relationship diagram, and the candidate conduction combinations, target current paths, and rectifier, inverter, and excitation coordinated actions are combined into candidate commutation control schemes. Based on the timing hybrid state model, the sequence of reachable state intervals of the current state set under each candidate commutation control scheme is calculated. When each reachable state interval in the sequence is located in the safe state domain and the last reachable state interval is located in the target current path state domain, a binding one-time valid zero-crossing identifier and the corresponding candidate commutation control scheme commutation permission information are generated. When both the one-time valid zero-crossing identifier and the commutation permission information are valid and unused, the candidate commutation control scheme bound to the commutation permission information is executed; When the actual state trajectory exceeds the corresponding reachable state interval sequence, or when the target current path state domain is not entered within the preset confirmation time window, the thyristor trigger pulse of the inverter bridge on the blocking machine side reduces the DC reactor current to a safe threshold and reverts to the commutation control state when the most recent commutation was successful.

Images