A railway tunnel dual backup power supply system switching and load balancing control method

Abstract

The application relates to the field of railway tunnel electromechanical engineering, and discloses a switching and load balancing control method for a railway tunnel dual backup power supply system, which comprises the following steps: acquiring power grid operation state parameters of a railway tunnel power supply section and a real-time power consumption sequence of load equipment; if any parameter in the power grid operation state parameters exceeds a preset threshold range, starting the dual backup power supply system; acquiring current environmental thermal state parameters and equipment operation parameters of the first generator set and the second generator set, and obtaining a first load carrying capacity index and a second load carrying capacity index; constructing a load demand matrix according to the real-time power consumption sequence; generating a load distribution strategy for the first generator set and the second generator set; and performing a switching operation of the load equipment according to the load distribution strategy, and adjusting the output power proportion of the two generator sets through closed-loop feedback during operation to realize load balancing control.

Description

Technical Field

[0001] This application relates to the field of railway tunnel electromechanical engineering, and in particular to a method for switching and load balancing control of a dual backup power supply system for railway tunnels. Background Technology

[0002] In railway construction under high-temperature environments, long tunnel complexes require a large number of electromechanical devices, including large jet fans, axial flow fans, water pumps, emergency lighting, and communication and signaling equipment. To cope with potential power outages from the municipal power grid, backup power systems consisting of large diesel generator sets are typically installed in the electromechanical control rooms at tunnel entrances and cross passages. To improve reliability, a design with dual backup power supplies operating in parallel is often adopted.

[0003] Traditional dual-standby power supply system switching and load distribution control methods are mostly based on pure electrical parameters for adjustment. The working logic is usually to start two generator sets simultaneously after detecting a mains power failure, and then connect them to the same bus after synchronizing the voltage and frequency. Relying on the droop control characteristics of the generator sets themselves or the active power sharing algorithm of the parallel controller, the two generator sets can bear the same load power.

[0004] However, the load-carrying capacity of generator sets will be greatly reduced under high temperature conditions. If the electrical load is still shared equally by two generator sets or distributed in a fixed ratio, once one of the generator sets is unexpectedly disconnected and shut down, the huge load it originally carried will be instantly transferred to the remaining single generator set. The huge inrush current can easily cause the remaining generator set to overload.

[0005] Solving this technical problem is a technical challenge that needs to be overcome by those skilled in the art. Summary of the Invention

[0006] This application provides a method for switching and load balancing control of a dual backup power supply system for railway tunnels, in order to at least partially solve the above-mentioned technical problems.

[0007] To achieve the above objectives, this application provides a method for switching and load balancing control of a dual backup power supply system for railway tunnels, comprising: Obtain the power grid operation status parameters and real-time power consumption sequence of the load equipment in the power supply section of the railway tunnel; If any parameter in the power grid operating status parameters exceeds a preset threshold range, the dual backup power system is activated; the dual backup power system includes a first generator set and a second generator set. Obtain the current environmental thermal state parameters and equipment operating parameters of the first and second generator sets, and obtain the first load-carrying capacity index and the second load-carrying capacity index. A load demand matrix is ​​constructed based on the real-time power consumption sequence; a load allocation strategy for the first generator set and the second generator set is generated based on the load demand matrix and the first load capacity index and the second load capacity index. The load distribution strategy is used to perform load switching operations and adjust the output power ratio of the two generator sets through closed-loop feedback during operation to achieve load balance control.

[0008] The embodiments of this application, through the above technical solution, solve the problem of unit overload caused by unreasonable load distribution in the dual backup power supply system of railway tunnels under high temperature environment, and improve the power supply reliability of the dual backup power supply system.

[0009] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating the steps of a method for switching and load balancing control of a dual backup power supply system for railway tunnels, provided in an exemplary embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the protection scope of this application.

[0013] This application provides a method for switching and load balancing control of a dual backup power supply system for railway tunnels. Please refer to [link / reference]. Figure 1 The switching and load balancing control method for a dual backup power supply system in a railway tunnel provided in this application includes the following steps: Step 101: Obtain the power grid operating status parameters and the real-time power consumption sequence of the load equipment in the railway tunnel power supply section. Specifically, the power grid operating status parameters include power grid voltage, frequency, phase, supply current, and power factor; the real-time power consumption sequence refers to the dataset formed by arranging the power consumption data generated by each electromechanical load equipment in the railway tunnel in chronological order.

[0014] Step 102: If any parameter in the power grid operating status parameters exceeds the preset threshold range, the dual backup power system is activated. The dual backup power system includes a first generator set and a second generator set. Specifically, the dual backup power system in this application consists of two large diesel generator sets, which are designed to operate in parallel. This design can provide continuous power to the jet fans, water pumps, emergency lighting, and communication signal equipment in the tunnel when the municipal power grid cannot provide normal power. The first generator set and the second generator set are the two core independent power supply units of the system. They can output power independently or jointly after being synchronized in parallel.

[0015] Step 103: Obtain the current environmental thermal state parameters and equipment operating parameters of the first and second generator sets, and obtain the first load-carrying capacity index and the second load-carrying capacity index. Specifically, the environmental thermal state parameters refer to parameters reflecting the thermal conditions of the generator set's operating environment and the unit's own thermal operating state, including the inlet temperature, outlet temperature, and multi-point temperature on the unit's surface; the equipment operating parameters refer to parameters reflecting the generator set's real-time operating conditions, including the unit's speed and excitation current; the load-carrying capacity index is an index used to characterize the generator set's current ability to withstand a step load. This index value is between zero and one, and the larger the value, the stronger the unit's current ability to withstand a step load. The first load-carrying capacity index is the load-carrying capacity corresponding to the first generator set, and the second load-carrying capacity index is the load-carrying capacity corresponding to the second generator set.

[0016] Step 104: Construct a load demand matrix based on the real-time power consumption sequence; generate a load allocation strategy for the first and second generator sets based on the load demand matrix and the first and second load capacity indicators. Specifically, the load demand matrix refers to a matrix formed by stacking the demand characteristic vectors of each load device in the railway tunnel; the load allocation strategy refers to a control strategy formulated based on the actual load capacity of the two generator sets and the power supply demand and priority of each load device, used to reasonably allocate the power supply task of each load device to the power supply branch corresponding to the first or second generator set. Unlike the traditional method of equally distributing load or allocating load in a fixed proportion, the load allocation strategy of this application matches the actual load capacity of the generator sets.

[0017] Step 105: Execute the switching operation of the load equipment according to the load distribution strategy, and adjust the output power ratio of the two generator sets through closed-loop feedback during operation to achieve load balance control. Specifically, the switching operation refers to the electrical operation of connecting or disconnecting the load equipment from the power supply branch of the generator set and connecting the generator set to the main power supply bus after the generator set meets the parallel operation conditions by controlling the closing or opening of the distribution circuit breaker and the main output circuit breaker of the generator set; closed-loop feedback adjustment refers to the adjustment method of monitoring various status parameters and operating indicators of the generator set during the operation of the dual standby power supply system and using the monitoring results as feedback signals to adjust the output power ratio of the two generator sets; load balance control is to match the load state of the two generator sets with their actual load capacity through the initial load distribution and the adjustment of the generator set output power, so as to avoid the overload operation of a single generator set.

[0018] Traditional control methods for dual-backup power supply systems in railway tunnels operating under high-temperature conditions rely solely on pure electrical parameter adjustments. When the generator's load-carrying capacity significantly decreases due to high temperatures, the system still distributes the load evenly or in a fixed proportion, posing a technical risk that the remaining generators might overload and shut down after a single generator is disconnected. This application addresses this issue by collecting grid operating status parameters and real-time load power consumption sequences to determine the start-up timing of the dual-backup power supply system, ensuring tunnel power supply. Based on generator environmental thermal state parameters and equipment operating parameters, a load-carrying capacity index reflecting actual operating conditions is obtained. A load demand matrix characterizing load demand is constructed based on the real-time load power consumption sequence, and a load allocation strategy is formulated in conjunction with the load-carrying capacity index, avoiding the drawbacks of fixed-proportion allocation. Load switching is executed according to the strategy, and dynamic load balance control is achieved through closed-loop feedback and dynamic adjustment of the generator output power ratio. Furthermore, the power allocation can be adjusted in real-time according to the generator's operating status, effectively mitigating the risk of single-generator overload. This solves the problem of generator overload caused by unreasonable load allocation in dual-backup power supply systems for railway tunnels operating under high-temperature conditions, improving the power supply reliability of the dual-backup power supply system.

[0019] In some embodiments, obtaining the current environmental thermal state parameters and equipment operating parameters of the first generator set and the second generator set, and obtaining the first load-carrying capacity index and the second load-carrying capacity index, includes: A feature vector is formed by sequentially concatenating environmental thermal state parameters and equipment operating parameters. The feature vector is a one-dimensional vector formed by integrating multiple generator set state parameters of different dimensions in a preset order, thus merging the thermal state and operating state of the unit into a unified feature representation.

[0020] The feature vector is input into a multi-layer feature extraction network; the multi-layer feature extraction network includes a first fully connected layer, a second fully connected layer, and an output layer. A multi-layer feature extraction network refers to a shallow neural network model composed of multiple fully connected layers and an output layer, used to extract and transform multi-dimensional feature vectors of generator sets layer by layer, uncovering the intrinsic correlation between different parameters. In this application, the network only has two levels of fully connected layers; a fully connected layer refers to a network layer in which the computational units of each layer are connected to all computational units of the previous layer, realizing linear transformation and non-linear mapping of features.

[0021] The feature vector is transformed through a first fully connected layer. This first fully connected layer is configured with the ReLU activation function and applies Dropout regularization to suppress overfitting, outputting the first hidden layer feature representation. The ReLU activation function is a non-linear activation function that introduces non-linearity into the neural network, addressing the problem of linear models failing to fit complex feature mapping relationships. Dropout regularization refers to a regularization method that randomly sets the output of some neurons to zero during feature processing in the neural network, effectively reducing co-adaptation relationships between neurons and improving the model's generalization ability. The first hidden layer feature representation refers to the feature data obtained after feature transformation and non-linear mapping by the first fully connected layer, representing a shallow abstraction of the generator set's original state parameters.

[0022] The first hidden layer feature representation is input into the second fully connected layer for higher-order feature extraction. The second fully connected layer contains a number of neurons and uses the ReLU activation function to output the second hidden layer feature representation. Higher-order feature extraction refers to further feature transformation and fusion of shallow abstract features to uncover the intrinsic relationships between different basic features, resulting in more representative abstract features that reflect the intrinsic connection between state parameters and the generator set's load-carrying capacity. Neurons are the basic computational units of a neural network, capable of receiving input signals and sequentially performing operations such as weighted summation and activation function transformation before outputting results. The number of neurons can be set according to the dimension of the generator set's state parameters and feature extraction requirements. The second hidden layer feature representation refers to the deep abstract features obtained after higher-order feature extraction by the second fully connected layer, reflecting the intrinsic relationship between the generator set's environmental thermal state, equipment operating parameters, and actual load-carrying capacity.

[0023] The output layer performs Softmax normalization on the feature representation of the second hidden layer, outputting a value between zero and one as the load-carrying capacity index of the corresponding generator set. A larger load-carrying capacity index indicates a stronger ability of the generator set to withstand a step load. Softmax normalization is a method that maps any real number output by the neural network to a value between 0 and 1, enabling the quantification and normalization of the generator set's load-carrying capacity, thus providing a unified comparison scale for the load-carrying capacity indices of different generator sets.

[0024] This application avoids the one-sidedness of using a single parameter to represent the unit's load-carrying capacity by concatenating environmental thermal state parameters and equipment operating parameters into a feature vector, making the feature representation more consistent with the actual operating state of the unit under high-temperature conditions. It performs layer-by-layer feature extraction through a multi-layer feature extraction network. The first fully connected layer uses Dropout regularization technology to effectively suppress model overfitting, ensuring the generalization ability of the load-carrying capacity index calculation model under different high-temperature conditions and different unit operating states. Through Softmax normalization, the deep abstract features are transformed into load-carrying capacity indices between 0 and 1, realizing the quantification of the unit's load-carrying capacity.

[0025] In some embodiments, constructing a load demand matrix based on a real-time power consumption sequence includes: The steady-state active power and peak starting power of each load device are extracted based on the real-time power consumption sequence. The duration of the transient impact on each load device from startup to steady state is calculated based on the decay trajectory of the real-time power consumption sequence. Steady-state active power refers to the continuous active power consumed by the load device after it completes the startup process and enters a stable operating state. It is the basic power parameter for normal operation and reflects the power demand when the device is working stably. The peak starting power refers to the maximum power consumption generated at the moment of startup of the load device. This value is usually much greater than the steady-state active power and can easily cause transient power impacts on the generator set. The decay trajectory of the real-time power consumption sequence refers to the continuous change curve of power consumption gradually decreasing from the peak starting power to the steady-state active power after the load device starts. The duration of the transient impact refers to the length of time from the start of the load device until the power consumption stabilizes at the steady-state active power, reflecting the continuous characteristics of the power impact on the generator set during the startup phase.

[0026] The safety assurance attributes of each load device in the tunnel electromechanical system are extracted, and corresponding power supply priority weights and maximum allowable switching delay times are assigned to each load device based on these attributes. Safety assurance attributes refer to the importance of load devices in the railway tunnel electromechanical system for tunnel operational safety, normal operation of core equipment, and emergency support; for example, jet fans that ensure personnel escape will be assigned higher priority weights and shorter delay times. Power supply priority weights are numerical values ​​that quantify the importance of power supply to load devices; a higher weight value indicates a higher power supply priority for the device, requiring priority processing during load allocation and switching. The maximum allowable switching delay time refers to the longest permissible time from the start-up of the dual backup power system to the completion of switching and obtaining stable power supply for the load device. Exceeding this time will affect the safe and stable operation of the tunnel electromechanical system; its value is determined by the device's safety assurance attributes.

[0027] The steady-state active power, peak starting power, transient impact duration, power supply priority weight, and maximum allowable switching delay time are concatenated to generate an independent demand feature vector for each load device. An independent demand feature vector is a one-dimensional vector formed by integrating five parameters characterizing the core power characteristics and power supply demand characteristics of a single load device in a preset order, representing all power supply demand information for that single load device.

[0028] A load demand matrix is ​​constructed by stacking the independent demand feature vectors of all load devices in rows. The rows of the load demand matrix are then sorted in descending order according to the product of the peak startup power and the power supply priority weight. Row stacking refers to the matrix construction method of arranging and combining the independent demand feature vectors of all load devices in rows to form a two-dimensional matrix. This method can integrate the scattered demand information of individual devices into a unified data structure. The product of the peak startup power and the power supply priority weight is the result of multiplying the two values. This result is used as the sorting criterion, taking into account both the startup impact degree and the importance of power supply for the load devices.

[0029] By extracting steady-state active power, peak starting power, and duration of transient impact from real-time power consumption sequences, the power characteristics of load equipment throughout its entire operating phase are captured, making the load demand representation more closely reflect the actual operating conditions of the equipment. By extracting safety assurance attributes and assigning corresponding power supply priority weights and maximum allowable switching delay times, the operational safety requirements of railway tunnels are integrated into the load demand representation, ensuring that core equipment with high safety assurance attributes is given priority in load allocation. Based on the product of the peak starting power and the power supply priority weight, the matrix row vectors are sorted in descending order, achieving a comprehensive sorting that takes into account both the degree of load impact and the importance of power supply. This allows subsequent load allocation to prioritize high-priority, high-impact load equipment, avoiding transient impacts on generator sets caused by their switching delays or concentrated switching.

[0030] In some embodiments, a load allocation strategy for the first generator set and the second generator set is generated based on the load demand matrix and a first load capacity index and a second load capacity index, including: Based on the first and second load-carrying capacity indicators, the current active power margin and transient impact tolerance threshold of the two generator sets are calculated. The active power margin refers to the maximum additional active power that the generator set can handle under current operating conditions. It is calculated by combining the unit's load-carrying capacity indicator with parameters such as rated active power and current actual output active power, reflecting the unit's remaining load-carrying potential. The transient impact tolerance threshold refers to the maximum value of the load equipment starting impact power that the generator set can withstand in a short period of time, obtained based on the unit's load-carrying capacity indicator.

[0031] Traverse the load demand matrix in descending order from top to bottom, extracting the target load device currently pointed to. If the first load capacity index is greater than the second load capacity index, the first generator set is set as the dominant unit; otherwise, the second generator set is set as the dominant unit. The dominant unit refers to the generator set with the higher load capacity index and stronger current step load capacity. During load allocation, it takes priority in powering the load device, reducing the operating pressure on the other generator set.

[0032] Compare the peak starting impact power of the target load equipment with the transient impact tolerance threshold of the dominant unit; if the transient impact tolerance threshold is greater than the peak starting impact power, then allocate the switching control command of the target load equipment to the power supply branch corresponding to the dominant unit and deduct the current active power margin of the dominant unit.

[0033] If the transient impact tolerance threshold is less than the peak starting impact power, the target load equipment is moved into the delayed waiting queue based on the maximum allowable switching delay time. Once the current active power margin of the dominant generator unit recovers to the preset threshold, the load equipment is allocated to the dominant generator unit sequentially according to the order of the delayed waiting queue. The remaining load equipment is weighted and allocated based on the ratio of the current active power margins of the two generator units. The delayed waiting queue is an ordered queue that temporarily stores load equipment that cannot be directly allocated due to insufficient transient impact tolerance threshold of the dominant generator unit. The queue maintains the original order of the load equipment in the load demand matrix, ensuring that high-priority loads are still processed first. Weighted allocation refers to allocating the remaining unallocated load equipment to the power supply branches of the two generator units according to the numerical ratio of their current active power margins, ensuring that the load allocation matches the actual remaining load capacity of the units.

[0034] This solution identifies the superior generating units and allocates loads differently according to demand, thereby adapting the unit's load capacity to the load power characteristics and power supply priority, effectively avoiding the risk of generator unit overload, and fully leveraging the overall efficiency of the two units.

[0035] In some embodiments, performing load switching operations on load devices according to a load distribution strategy includes: The system monitors the output electrical parameters of the first and second generator sets in real time. When the voltage, frequency, and phase difference all meet the preset synchronous parallel operation conditions, the parallel operation control cabinet closes the main output circuit breakers of the first and second generator sets to establish the main power supply bus voltage. Synchronous parallel operation conditions refer to the electrical parameter matching threshold requirements that the two generator sets must meet to achieve parallel operation. The parallel operation control cabinet is a dedicated control device integrating generator set parallel operation control, electrical parameter monitoring, and circuit breaker on / off control functions; it is the core control unit for the parallel operation of the dual backup power supply system. The main output circuit breaker is an on / off control switch installed between the generator set output and the main power supply bus, used to control the connection and disconnection of the generator set output power to the main power supply bus. The main power supply bus voltage refers to the stable operating voltage required for the main power supply bus to supply power to all load equipment in the tunnel after the two generator sets are paralleled.

[0036] The load devices in the first row of the load demand matrix are extracted and sorted. A closing action is triggered within the basic time window after the main power supply bus voltage is established. The basic time window refers to the shortest waiting time set after the main power supply bus voltage is established to ensure the bus voltage stabilizes, avoiding secondary voltage fluctuations caused by switching loads before the bus voltage is stable. The closing action refers to controlling the circuit breaker of the corresponding distribution circuit of the load device to complete the closing operation, establishing an electrical connection between the load device and the main power supply bus.

[0037] During the execution of the closing action, the transient voltage drop amplitude of the main power supply bus is monitored. If the transient voltage drop amplitude exceeds the set safety boundary, a transient blocking mechanism is triggered. The transient blocking mechanism includes freezing and delaying the closing actions of circuit breakers for subsequent load devices in the waiting queue. The transient voltage drop amplitude refers to the specific value of the instantaneous voltage drop of the main power supply bus during the switching of load devices, reflecting the impact of the load switching operation on the stability of the bus voltage. The safety boundary refers to the maximum allowable value of the transient voltage drop amplitude preset to ensure the safe operation of the main power supply bus and tunnel electromechanical equipment. If this value is exceeded, the bus voltage is determined to be in an unsafe operating state. The transient blocking mechanism is a protection mechanism that temporarily blocks subsequent load switching operations when the main power supply bus voltage experiences an over-threshold drop to prevent further voltage deterioration.

[0038] Once the transient voltage drop amplitude of the main power supply bus recovers and remains within the stable operating range, the transient blocking mechanism is released. Following the time interval command set by the maximum permissible switching delay time, the remaining distribution circuit breakers for each load device are sequentially activated and closed. The stable operating range refers to the main power supply bus voltage range that ensures the normal operation of all load devices in the tunnel; it is the basis for determining whether the bus voltage has stabilized and load switching can continue. Distribution circuit breakers are independent on / off control switches installed on the distribution circuits of each load device, used to control the connection and disconnection between a single load device and the main power supply bus.

[0039] This solution monitors the bus voltage in real time and triggers a transient blocking mechanism to prevent voltage instability. It also switches the remaining loads in an orderly manner at time intervals, thus achieving safe and orderly switching of load equipment and effectively ensuring the stable operation of the main power supply bus of the dual backup power system.

[0040] In some embodiments, load balancing control is achieved by adjusting the output power ratio of the two generator sets through closed-loop feedback during operation, including: The system monitors the rate of change of environmental thermal state parameters and the current active power output of the first and second generator sets. When the surface temperature change rate of the first generator set continuously exceeds the preset warning value for a set duration and the first load-bearing capacity index shows a downward trend, the target power transfer amount required to restore thermal balance is calculated based on the thermal deviation value of the first generator set. The rate of change of environmental thermal state parameters refers to the change value of the inlet temperature, outlet temperature, and multi-point temperature on the surface of the generator set per unit time, reflecting the real-time trend of the generator set's thermal state. The preset warning value is a critical value of the rate of change of environmental thermal state parameters pre-calibrated to ensure the safe operation of the generator set. Exceeding this value indicates that the thermal state of the unit is developing in an abnormal direction. The thermal deviation value is the difference between the current actual thermal state parameters of the generator set and the thermal state parameters during normal stable operation. The target power transfer amount is the specific value of active power that needs to be unloaded from the generator set and transferred to another generator set to restore the generator set with abnormal thermal state to a thermal balance state.

[0041] Obtain the current active power margin of the second generator set and the rate of change of its environmental thermal state parameters, and calculate the dynamic safe receiving limit of the second generator set. The dynamic safe receiving limit refers to the maximum active power that the second generator set can safely receive from the first generator set under the current thermal state and power operating conditions, and is calculated by combining its current active power margin and the rate of change of its environmental thermal state parameters.

[0042] The smaller value between the target transferred power and the dynamic safe receiving limit is taken as the actual transferred power, and a power transfer command carrying the actual transferred power and a preset slope is generated. The actual transferred power refers to the actual active power value for cross-unit transfer determined after comprehensively considering the power unloading demand of the thermally abnormal unit and the safe receiving capacity of the other unit; the preset slope refers to the rate limit value when the active power output of the generator set is adjusted. Setting a preset slope can avoid electrical shocks to the power supply system caused by sudden power output changes; the power transfer command is a control command carrying the actual transferred power and the preset slope to control the coordinated adjustment of the active power output ratio of the two generator sets.

[0043] Based on the power transfer command, the excitation systems and electronic speed governors of the two generator sets are coordinated to smoothly reduce the active power share of the first generator set at a preset slope, while simultaneously increasing the active power share of the second generator set by the same amount to ensure that the total output power matches the real-time power consumption sequence. The excitation system refers to the system in the generator set used to adjust the excitation current, thereby controlling the generator output voltage and active power; the electronic speed governor is a device that dynamically adjusts the generator set speed and active power output by regulating the fuel supply; the active power share refers to the proportion of the active power output of a single generator set to the total active power output of the two generator sets.

[0044] This solution achieves dynamic closed-loop balance control of generator power, effectively avoiding thermal overload of the unit, ensuring smooth and shock-free power regulation, and matching the total system output power with the real-time load demand.

[0045] In some embodiments, the method further includes: After a preset time window for coordinating the output power ratio of the two generator sets, the rate of change of environmental thermal state parameters of the first and second generator sets is monitored. The preset time window refers to a fixed monitoring period reserved after the coordinated adjustment of the generator set output power ratio is completed, in order to observe the effect of power adjustment on the improvement of the unit's thermal state.

[0046] When the rate of change of the environmental thermal state parameters of both the first and second generator sets remains higher than the set warning value, power distribution circuits with power supply priority weights lower than a preset threshold are extracted as candidate load devices based on the load demand matrix. Candidate load devices refer to load devices with lower power supply priority selected from the tunnel electromechanical system, whose temporary unloading will not affect the core safety of tunnel operation.

[0047] Collect the operating power factor and reactive power consumption of each candidate load device. The operating power factor is the ratio of active power to apparent power when the candidate load device is running, reflecting the energy utilization efficiency of the device; the reactive power consumption is the amount of reactive power consumed by the candidate load device during operation.

[0048] The unloading priority index for each candidate load device is calculated based on reactive power consumption, operating power factor, and power supply priority weight. Reactive power consumption is positively correlated with the unloading priority index, while operating power factor and power supply priority weight are negatively correlated with the unloading priority index. The unloading priority index is an indicator of the unloading priority obtained by comprehensively considering the reactive power consumption, energy utilization efficiency, and power supply importance of the candidate load device. The higher the index, the higher the priority that the device should be unloaded.

[0049] Candidate load devices are sorted in descending order of their offloading priority index to construct an offloading queue. An offloading queue is an ordered queue formed by arranging candidate load devices in descending order of their offloading priority index.

[0050] The power distribution circuits corresponding to each candidate load device in the unloading queue are disconnected sequentially, and the cooling fan speed of the generator set is increased to the highest priority. The highest priority of cooling fan speed means that the operating speed of the generator set cooling fan is adjusted to the maximum speed designed for the equipment, so as to maximize the heat dissipation efficiency of the generator set and speed up the recovery of the unit's thermal state to the normal operating range.

[0051] This solution addresses the lack of targeted unloading strategies when both generator sets experience thermal overload in traditional control systems. It can quickly reduce the thermal load on both generator sets and effectively avoid shutdown failures caused by continuous thermal overload.

[0052] In some embodiments, the method further includes: During the closing operation, the transient zero-sequence potential fluctuation amplitude of the grounding grid is collected by a zero-sequence voltage sensor deployed in the grounding grid of the tunnel electromechanical control room. The zero-sequence voltage sensor is a sensing device specifically used to detect the zero-sequence voltage signal of the power system and capture transient changes in the grounding grid potential; the transient zero-sequence potential fluctuation amplitude refers to the magnitude of the instantaneous fluctuation of the zero-sequence potential of the grounding grid in the tunnel electromechanical control room during the electrical operation of load switching, reflecting the potential stability state of the grounding grid.

[0053] When the amplitude of the transient zero-sequence potential fluctuation exceeds the preset safety threshold and the duration exceeds the preset hysteresis window, it is determined that there is a current dissipation hysteresis effect caused by low conductivity geology, and the release condition of the transient blocking mechanism is updated to a dual verification of the bus voltage and the grounding grid potential. Low conductivity geology refers to the geological conditions in which the soil and other geological media in the railway tunnel area have poor conductivity, which will hinder the current dissipation process in the grounding grid; the current dissipation hysteresis effect refers to the phenomenon that due to the low geological conductivity, the impulse current or fault current in the grounding grid cannot be dissipated to the ground quickly and smoothly, thus causing the zero-sequence potential of the grounding grid to remain in an abnormal state; dual verification means that the stable state of the main power supply bus voltage and the stable state of the grounding grid zero-sequence potential are used together as the release judgment condition of the transient blocking mechanism, and the release command can only be executed if both conditions are met.

[0054] The instruction to release the transient blocking mechanism will be executed when the following two conditions are met simultaneously: the transient voltage drop amplitude of the main power supply bus falls back to the preset voltage threshold and is maintained for a preset duration, and the transient zero-sequence potential fluctuation amplitude falls back to the preset safety threshold and is maintained for a preset stable duration.

[0055] The time compensation is calculated based on the actual hysteresis time taken for the transient zero-sequence potential fluctuation amplitude to fall from its peak to a preset safety threshold. The time compensation is the product of a preset adjustment coefficient and the actual hysteresis time. The actual hysteresis time refers to the actual time it takes for the transient zero-sequence potential fluctuation amplitude of the grounding grid to decrease from its peak to the preset safety threshold, reflecting the degree of influence of the current-dissipating hysteresis effect on the grounding grid potential recovery. The time compensation is a specific time value calculated based on the actual hysteresis time, used to extend the maximum allowable switching delay time of the load equipment. The preset adjustment coefficient is a fixed coefficient pre-calibrated based on the geological characteristics of the railway tunnel and the operating parameters of the power supply system, used to determine the proportional relationship between the actual hysteresis time and the time compensation.

[0056] By dynamically extending the maximum allowable switching delay time of each load device in the delay waiting queue using time compensation, a new switching scheduling sequence is formed. The new switching scheduling sequence refers to the load switching order and timing arrangement adapted to the abnormal grounding grid potential state, which is formed on the basis of the original load switching scheduling arrangement by dynamically extending the maximum allowable switching delay time of each load device in the delay waiting queue.

[0057] This solution addresses the problems of traditional load switching that only monitor bus voltage, neglecting the current dissipation hysteresis effect and potential anomalies caused by low-conductivity geological conditions. It also addresses the issues of premature unlocking leading to instability in both the power supply system and the grounding grid, and the lack of a dynamic adjustment mechanism for load switching adapted to geological characteristics. By monitoring the grounding grid potential to determine the current dissipation hysteresis effect, the solution optimizes the unlocking process to a dual verification of bus voltage and grounding grid potential, ensuring the stability of both the power supply system and the grounding grid. The solution calculates time compensation to adjust the load switching sequence, ensuring that load switching adapts to potential anomalies caused by geological characteristics.

[0058] In some embodiments, the environmental thermal state parameters include the inlet temperature, the outlet temperature, and the multi-point temperature on the surface of the generator set, and the equipment operating parameters include the generator set speed and the excitation current.

[0059] In some embodiments, in the calculation of the unloading priority index, reactive power consumption is positively correlated with the unloading priority index, while the operating power factor and power supply priority weight are negatively correlated with the unloading priority index.

[0060] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0061] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0062] The embodiments, implementation methods, and related technical features of this application can be combined and substituted for each other without conflict.

[0063] The above are merely preferred embodiments of this application and are not intended to limit this application in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application without departing from the scope of the technical solution of this application shall still fall within the scope of the technical solution of this application.

Claims

1. A method for switching and load balancing control of a dual backup power supply system for railway tunnels, characterized in that, include: Obtain the power grid operation status parameters and real-time power consumption sequence of the load equipment in the power supply section of the railway tunnel; If any parameter in the power grid operating status parameters exceeds a preset threshold range, the dual backup power system is activated; the dual backup power system includes a first generator set and a second generator set. Obtain the current environmental thermal state parameters and equipment operating parameters of the first and second generator sets, and obtain the first load-carrying capacity index and the second load-carrying capacity index. A load demand matrix is ​​constructed based on the real-time power consumption sequence; a load allocation strategy for the first generator set and the second generator set is generated based on the load demand matrix and the first load capacity index and the second load capacity index. The load distribution strategy is used to perform load switching operations and adjust the output power ratio of the two generator sets through closed-loop feedback during operation to achieve load balance control.

2. The method according to claim 1, characterized in that, Obtain the current environmental thermal state parameters and equipment operating parameters of the first and second generator sets, and obtain the first load-carrying capacity index and the second load-carrying capacity index, including: The environmental thermal state parameters and equipment operating parameters are sequentially concatenated to form a feature vector; The feature vector is input into a multi-layer feature extraction network; the multi-layer feature extraction network includes a first fully connected layer, a second fully connected layer, and an output layer. The feature vector is transformed by the first fully connected layer; wherein the first fully connected layer is configured with the ReLU activation function and applies Dropout regularization to suppress overfitting, and outputs the first hidden layer feature representation; The first hidden layer feature representation is input into the second fully connected layer for high-order feature extraction; wherein the second fully connected layer contains a number of neurons and uses the ReLU activation function to output the second hidden layer feature representation. The second hidden layer feature representation is normalized by Softmax through the output layer, and the output value between zero and one is used as the load capacity index of the corresponding generator set; the larger the load capacity index value, the stronger the ability of the generator set to withstand step load.

3. The method according to claim 2, characterized in that, Constructing a load demand matrix based on the real-time power consumption sequence includes: Based on the real-time power consumption sequence, the steady-state operating active power and the peak value of the starting impact power of each load device are extracted, and the transient impact duration of each load device from startup to entering steady state is calculated according to the decay trajectory of the real-time power consumption sequence. Extract the safety assurance attributes of each load device in the tunnel electromechanical system, and assign corresponding power supply priority weights and maximum allowable switching delay times to each load device according to the safety assurance attributes; The steady-state operating active power, the peak value of the starting impact power, the duration of the transient impact, the power supply priority weight, and the maximum allowable switching delay time are concatenated to generate an independent demand feature vector for each load device. The load demand matrix is ​​constructed by stacking the independent demand feature vectors of all load devices in rows, and the row vectors of the load demand matrix are sorted in descending order according to the product of the peak startup power and the power supply priority weight.

4. The method according to claim 3, characterized in that, Based on the load demand matrix and the first and second load capacity indices, a load allocation strategy is generated for the first and second generator sets, including: Based on the first load capacity index and the second load capacity index, calculate the current active power margin and transient impact tolerance threshold of the two generator sets. Traverse the load demand matrix in descending order from top to bottom, and extract the target load device currently pointed to; if the first load capacity index is greater than the second load capacity index, then set the first generator set as the dominant generator set, otherwise set the second generator set as the dominant generator set. Compare the peak starting impact power of the target load device with the transient impact tolerance threshold of the dominant unit; if the transient impact tolerance threshold is greater than the peak starting impact power, then allocate the switching control command of the target load device to the power supply branch corresponding to the dominant unit and deduct the current active power margin of the dominant unit. If the transient impact tolerance threshold is less than the peak value of the starting impact power, the target load device is moved into the delay waiting queue based on the maximum allowable switching delay time. When the current active power margin of the dominant unit recovers to the preset threshold, the load devices are allocated to the dominant unit in sequence according to the order of the delay waiting queue. The remaining load devices are weighted and allocated according to the ratio of the current active power margin of the two generator sets.

5. The method according to claim 4, characterized in that, Performing load switching operations on load devices according to the load distribution strategy includes: Real-time monitoring of the output electrical parameters of the first and second generator sets; when the voltage, frequency and phase difference all meet the preset synchronous parallel operation conditions, control the parallel operation control cabinet to close the main output circuit breakers of the first and second generator sets to establish the main power supply bus voltage of the system. Extract the first row of load devices after sorting the load demand matrix, and trigger a closing action within the basic time window after the main power supply bus voltage is established; During the execution of the closing action, the transient voltage drop amplitude of the main power supply bus is monitored. If the transient voltage drop amplitude exceeds the set safety boundary, a transient blocking mechanism is triggered. The transient blocking mechanism includes freezing the circuit breaker closing actions of subsequent load devices in the delayed waiting queue. Once the transient voltage drop amplitude of the main power supply bus recovers and remains within the stable operating range, the transient blocking mechanism is released, and the circuit breakers of the remaining load equipment are activated and closed sequentially according to the time interval command set by the maximum allowable switching delay time.

6. The method according to claim 5, characterized in that, During operation, load balance control is achieved by adjusting the output power ratio of the two generator sets through closed-loop feedback, including: Monitor the rate of change of environmental thermal state parameters and the current active power output value of the first generator set and the second generator set; when the surface temperature change rate of the first generator set continues to exceed the preset warning value for a preset time and the first load capacity index shows a downward trend, calculate the target transfer power required to restore thermal balance based on the thermal deviation value of the first generator set. Obtain the current active power margin of the second generator set and the rate of change of its environmental thermal state parameters, and calculate the dynamic safe reception limit of the second generator set; The smaller value between the target transfer power and the dynamic secure reception limit is taken as the actual transfer power, and a power transfer command carrying the actual transfer power and a preset slope is generated. Based on the power transfer command, the excitation systems and electronic speed governors of the two generator sets are coordinated to smoothly reduce the output active power share of the first generator set at the preset slope and simultaneously increase the output active power share of the second generator set by the same amount to ensure that the total output power matches the real-time power consumption sequence.

7. The method according to claim 6, characterized in that, The method further includes: After coordinating and adjusting the output power ratio of the two generator sets for a preset time window, monitor the rate of change of the environmental thermal state parameters of the first and second generator sets. When the rate of change of the environmental thermal state parameters of the first generator set and the second generator set is continuously higher than the set warning value, the power distribution circuits with power supply priority weights lower than the preset threshold are extracted as candidate load devices based on the load demand matrix. Collect the operating power factor and reactive power consumption of each candidate load device; The offloading priority index of each candidate load device is calculated based on reactive power consumption, operating power factor, and power supply priority weight. The candidate load devices are sorted in descending order of the unloading priority index to construct an unloading queue; The power distribution circuits corresponding to each candidate load device in the unloading queue are disconnected in sequence, and the cooling fan speed of the generator set is increased to the highest priority.

8. The method according to claim 7, characterized in that, The method further includes: During the execution of the closing action, the transient zero-sequence potential fluctuation amplitude of the grounding grid is collected by a zero-sequence voltage sensor deployed in the grounding grid of the tunnel electromechanical control room; When the amplitude of the transient zero-sequence potential fluctuation exceeds the preset safety threshold and the duration exceeds the preset hysteresis window, it is determined that there is a current hysteresis effect caused by low conductivity geology, and the release condition of the transient blocking mechanism is updated to a dual verification of bus voltage and grounding grid potential. When the following two conditions are met simultaneously, the instruction to release the transient blocking mechanism is executed: the transient voltage drop amplitude of the main power supply bus falls back to the preset voltage threshold and is maintained for a preset duration, and the transient zero-sequence potential fluctuation amplitude falls back to the preset safety threshold and is maintained for a preset stability duration. The time compensation amount is calculated based on the actual hysteresis time taken for the transient zero-sequence potential fluctuation amplitude to fall from its peak to the preset safety threshold; wherein, the time compensation amount is the product of the preset adjustment coefficient and the actual hysteresis time; The maximum allowable switching delay time of each subsequent load device in the delay waiting queue is dynamically extended using the time compensation amount to form a new switching scheduling sequence.

9. The method according to claim 8, characterized in that, The environmental thermal state parameters include the air inlet temperature, air outlet temperature, and multi-point temperature on the surface of the generator set. The equipment operating parameters include the generator set speed and excitation current.

10. The method according to claim 9, characterized in that, In the step of calculating the unloading priority index of each candidate load device based on reactive power consumption, operating power factor and power supply priority weight, reactive power consumption is positively correlated with unloading priority index, while operating power factor and power supply priority weight are negatively correlated with unloading priority index.

Images