A substation external emergency power source automatic throw-in system reconstruction method
By constructing an emergency power supply network and neural network model, the problems of untimely and erroneous operation of traditional substation backup power automatic transfer systems have been solved, enabling rapid and accurate activation of substation backup power and improving the reliability and stability of power supply.
Patent Information
- Application Number
- CN202511576694.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-31
- Publication Date
- 2026-02-10
- Estimated Expiration
- 2045-10-31
AI Technical Summary
Traditional substation backup power automatic transfer systems rely on manual operation and simple logic judgment, resulting in untimely backup power supply and high risk of misoperation, making it difficult to meet the reliability and stability requirements of modern substation power supply.
An emergency power supply network is constructed, electrically operated circuit breakers are configured, and convolutional neural networks and long short-term memory network models are combined to achieve automated control and accurate fault diagnosis, thereby optimizing the process of switching on backup power.
It enables the rapid and accurate deployment of backup power in substations, improves the reliability and stability of power supply, shortens power outage time, and enhances the ability to identify and handle complex faults.
Smart Images

Figure CN121055558B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power supply system technology, and more specifically, to a method for modifying an external emergency power supply automatic transfer system for substations. Background Technology
[0002] In the operation of power systems, the reliability and stability of substations are crucial, and the rapid and accurate activation of backup power is one of the key aspects of ensuring the continuity of power supply from substations. Traditional automatic transfer switching systems for substation backup power rely heavily on simple logical judgments and manual operation, which has many limitations.
[0003] Specifically, traditional systems involve numerous manual operations, which are not only inefficient but also prone to human error leading to backup power failures or uncoordinated actions, further increasing the risk of substation power outages. For example, in a conventional, older 110kV substation, all 10kV circuit breakers are manually operated for switching modes, and the 10kV external emergency power supply is also manually switched on. When any of the following situations occur in the 110kV substation power supply: loss of power to two 110kV external lines, 110kV GIS equipment failure, failure of two 110kV main transformers, or failure of two 10kV incoming line switches, and the failure cannot be repaired quickly, manual operation of the substation's 10kV switchgear is required. A typical 110kV substation has 60 10kV switchgear panels, while a large 110kV substation has more than 80. Due to the numerous bays in the 10kV switchgear and the complex and cumbersome switching procedures according to the "State Grid Corporation of China Power Safety Work Regulations," the switching and activation of the 10kV emergency power supply takes a long time, resulting in a prolonged power outage for the entire substation. This prolonged power outage for all users will lead to significant property damage and social impact. Furthermore, given the high reliability requirements of substation operation, simply configuring manual operation to automatic control carries a significant risk of misoperation. Therefore, it is necessary to design an automated logic operation that accurately and reliably executes the emergency power supply activation logic in emergency situations, thereby reducing outage time and quickly restoring power supply.
[0004] Furthermore, traditional systems suffer from delayed fault detection, typically relying on threshold judgments based on basic parameters such as voltage and current. Backup power is only activated when these parameters reach a certain level of abnormality. This delay in backup power activation can hinder timely fault response, prolonging outages and impacting power supply reliability. Secondly, traditional systems have limited fault diagnosis accuracy and insufficient ability to identify complex fault scenarios. Substation faults are diverse, exhibiting not only different voltage and current variations but also potentially involving changes in multiple switching states. Traditional systems struggle to synthesize this complex information for comprehensive and accurate fault diagnosis, leading to misdiagnosis or missed diagnoses, thus affecting the correctness and effectiveness of backup power activation.
[0005] With the continuous expansion of power systems and the increasing demands for power supply reliability, traditional automatic transfer switching (ATS) systems for backup power supplies are no longer sufficient to meet the operational needs of modern substations. In recent years, the rapid development of artificial intelligence technology and its increasingly widespread application in various fields have brought new opportunities to ATS systems for substation backup power supplies.
[0006] Artificial intelligence (AI) technology, especially neural network models, possesses powerful data processing and pattern recognition capabilities. It can perform in-depth analysis and mining of massive amounts of substation operational data, enabling early warning and accurate fault assessment, and providing strong support for the rapid and correct deployment of backup power. However, in the field of automatic backup power transfer (AS / RS) in substations, solutions that truly integrate AI technology with traditional AS / RS systems are still relatively rare. Most existing technologies still rely primarily on traditional logical judgment and manual operation, failing to fully leverage the advantages of AI to address problems in the backup power deployment process. Therefore, how to combine modern AI technology to transform and optimize substation AS / RS systems to improve the timeliness and accuracy of fault detection, optimize the backup power deployment process, and ensure the continuity and stability of substation power supply has become a pressing technical problem in this field. Summary of the Invention
[0007] The present invention aims to solve at least one of the aforementioned technical problems existing in the prior art.
[0008] Therefore, the present invention provides a method for modifying an external emergency power supply automatic transfer system in a substation.
[0009] This invention provides a method for retrofitting an external emergency power supply automatic transfer system in a substation, comprising:
[0010] Construct an emergency power supply network; wherein, the substation includes a primary bus and a secondary bus, and a power supply link is constructed between the emergency power station and the secondary bus of the substation, and the power supply voltage level of the emergency power station and the secondary bus of the substation is the same.
[0011] The core circuit breaker is configured as an electrically operated circuit breaker, which performs on / off operations according to control signals; the core circuit breaker is a circuit breaker related to the automatic transfer of emergency power.
[0012] Collect substation operation data, which includes at least voltage data, current data, protection interlocking signals, and switch information at the collection points;
[0013] Based on the substation's operating data, determine whether the current operating status of the substation requires the activation of emergency power supply;
[0014] When the current operating status of the substation indicates that emergency power supply needs to be activated, the system determines whether the emergency power automatic transfer switching conditions are met based on the real-time operating data of the substation's external emergency power automatic transfer system. If the conditions are met, a control signal is sent to each core circuit breaker. The control signal controls each core circuit breaker to supply emergency power to the substation's secondary bus according to the automatic transfer switch operation control logic.
[0015] The substation external emergency power supply automatic transfer system modification method according to the above-described technical solution of the present invention may also have the following additional technical features:
[0016] In the above technical solution, the secondary side of the substation has several busbar sections, and the emergency power station has N busbar sections; the construction of the emergency power supply network includes:
[0017] The substation secondary side busbars are divided into N groups, each corresponding to a busbar in the emergency power station. Each busbar in each group on the secondary side of the substation is connected to the corresponding busbar in the emergency power station via an independent transmission line. Corresponding substation-side incoming circuit breakers and emergency power station-side outgoing circuit breakers are installed on the transmission lines.
[0018] In the above technical solution, at least one main transformer is installed between the primary busbar and the secondary busbar of the substation, and the secondary side of each main transformer is connected to the corresponding secondary busbar section through a main transformer circuit breaker.
[0019] Emergency power supply is only permitted to be activated when all busbar sections on the secondary side of the substation are de-energized.
[0020] Fault types that can cause all busbar sections on the secondary side of a substation to lose power include: loss of power to all incoming lines on the primary side of the substation, failure of all main transformers in the substation, and failure of all main transformer circuit breakers in the substation.
[0021] In the above technical solution, determining whether the current operating status of the substation requires the activation of emergency power supply based on substation operating data includes:
[0022] The decision on whether to activate emergency power supply is made directly based on the comparison between the substation's real-time operating data and the system's preset values.
[0023] Alternatively, based on historical and real-time operating data of the substation, a neural network model can be used to predict or detect substation faults, and the need to activate emergency power supply can be determined based on the fault situation.
[0024] In the above technical solution, the neural network model includes a convolutional neural network and a long short-term memory network;
[0025] The method of using a neural network model to predict or detect substation faults includes:
[0026] Obtain historical operating data of substations arranged in chronological order and their corresponding fault information;
[0027] Preprocessing of historical operating data of substations includes data cleaning and normalization.
[0028] Feature extraction is performed on the preprocessed historical operation data of the substation, and combined with the corresponding fault condition labels to form an operation dataset;
[0029] The neural network model is trained based on the running dataset; wherein, the Adam optimizer and cross-entropy loss function are used to train the neural network model to obtain the trained model;
[0030] Based on real-time operating data of the substation, a trained neural network model is used to determine the current fault type of the substation.
[0031] In the above technical solution, after determining the current fault type of the substation based on the neural network model, if the current fault type of the substation will cause all busbar sections on the secondary side of the substation to lose voltage, it is determined that emergency power supply needs to be started.
[0032] In the above technical solution, the network architecture of the neural network model includes:
[0033] The input layer receives preprocessed multidimensional time series data.
[0034] Convolutional layers use one-dimensional convolution to extract local features from multi-dimensional time series data;
[0035] LSTM layers are used to capture long-term dependencies in time series data;
[0036] Fully connected layers are used to integrate extracted local features;
[0037] The output layer uses the softmax activation function to output the probability distribution of fault types; when the probability of a certain fault type in the output structure exceeds a set threshold, the system is considered to have encountered that fault.
[0038] In the above technical solution, when the neural network model predicts that the substation fault type will cause all busbar sections on the secondary side of the substation to lose voltage at the next moment, the emergency power supply is put into hot standby state before the emergency power supply is started.
[0039] In the hot standby state, the outgoing circuit breaker on the emergency power station side is in operation, and the incoming circuit breaker on the substation side is in hot standby state.
[0040] In the above technical solution, the criteria for determining whether the emergency power automatic transfer conditions are met based on real-time operating data of the substation's external emergency power automatic transfer system include:
[0041] All busbar sections on the secondary side of the substation lost power;
[0042] There is no current in the circuit where the main transformer circuit breaker is located;
[0043] The circuit breaker of the main transformer is located on a line without a protection lockout signal.
[0044] The busbar section of the emergency power station currently scheduled to be put into operation has been charged. Specifically, if the secondary busbar section of the substation was energized before the fault occurred, the outgoing circuit breaker of the emergency power station was in operation, the incoming circuit breaker of the substation was in cold standby or hot standby status, and the busbar section of the emergency power station was energized for a preset period of time, it is determined that the busbar section of the emergency power station currently scheduled to be put into operation has been charged.
[0045] In the above technical solution, the automatic switching action control logic includes:
[0046] Disconnect all feeder switches corresponding to the secondary busbar section of the substation.
[0047] Disconnect the main transformer circuit breaker corresponding to the secondary busbar section of the substation.
[0048] Control the closing of the substation-side incoming circuit breaker corresponding to the secondary busbar section of the substation.
[0049] After the secondary busbar section of the substation is energized, at least some of the corresponding feeder switches are closed in sequence.
[0050] In summary, due to the adoption of the above-mentioned technical features, the beneficial effects of the present invention are:
[0051] This invention constructs an emergency power supply network, configures core circuit breakers as electrically operated circuit breakers, and combines automated control logic and advanced neural network models to achieve rapid and accurate deployment of backup power in substations. This significantly improves the reliability and stability of substation power supply, effectively shortens outage time, accelerates power restoration, and enhances the system's ability to identify and handle complex faults, providing strong support for the stable operation of the power system.
[0052] Specifically, firstly, in terms of automation control, this invention achieves automated control of the backup power supply process by using electrically operated circuit breakers to perform on / off operations based on control signals. This automated control method can respond quickly to fault situations, significantly shortening power outage time. When a fault occurs, the system can quickly determine the need to activate emergency power supply and rapidly send control signals to each core circuit breaker, enabling the emergency power supply to supply power to the secondary busbar of the substation, thereby accelerating the restoration of power and reducing the impact of power outages on users.
[0053] Secondly, the control logic of this invention has a significant impact on system reliability. When determining whether the emergency power supply automatic switching conditions are met, multiple factors are comprehensively considered, including the undervoltage status of all busbar sections on the secondary side of the substation, the current and protection blocking signals of the line where the main transformer circuit breaker is located, and the charging status of the emergency power supply station busbar sections. This comprehensive and rigorous control logic ensures accurate activation of the backup power supply while meeting safety and stability requirements, avoiding incorrect activation or refusal of the backup power supply due to misjudgment, thereby improving the overall system reliability.
[0054] Finally, the neural network model employed in this invention has significant advantages. This model combines Convolutional Neural Networks (CNNs) and Long Short-Term Memory Networks (LSTMs), enabling in-depth analysis and mining of substation operational data. CNNs can extract local features from multi-dimensional time-series data, while LSTMs can capture long-term dependencies in the time series. In this way, the neural network model can more accurately predict and detect substation faults, anticipating their occurrence and providing more time for backup power supply deployment. Furthermore, the neural network model can handle complex nonlinear relationships, improving the accuracy of fault diagnosis and further enhancing the system's intelligence and ability to handle complex faults.
[0055] Additional aspects and advantages of the invention will become apparent in the following description or may be learned by practice of the invention. Attached Figure Description
[0056] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:
[0057] Figure 1 This is a flowchart of a method for modifying an external emergency power supply automatic transfer system in a substation according to an embodiment of the present invention;
[0058] Figure 2 This is a system architecture diagram of the substation external emergency power automatic transfer system in an embodiment of the present invention;
[0059] Figure 3 This is a flowchart illustrating the process of using a neural network model to predict or detect substation faults in a substation external emergency power supply automatic transfer system modification method according to an embodiment of the present invention.
[0060] Figure 4 This is a neural network model architecture diagram used in the substation external emergency power supply automatic transfer system modification method according to an embodiment of the present invention;
[0061] Figure 5 This is a schematic diagram of the emergency power automatic transfer control logic of the 10kV I busbar in a substation external emergency power automatic transfer system modification method according to an embodiment of the present invention. Detailed Implementation
[0062] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.
[0063] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.
[0064] The following reference Figures 1 to 5 This invention describes a method for retrofitting an external emergency power supply automatic transfer system for substations, provided by some embodiments of the present invention.
[0065] Some embodiments of this application provide a method for retrofitting an external emergency power supply automatic transfer system for a substation.
[0066] like Figure 1 As shown, the first embodiment of the present invention proposes a method for modifying an external emergency power supply automatic transfer system in a substation, including the following steps S1 to S5.
[0067] S1. Construct an emergency power supply network; wherein, the substation includes a primary bus and a secondary bus, and a power supply link is constructed between the emergency power station and the secondary bus of the substation, and the power supply voltage level of the emergency power station and the secondary bus of the substation is the same.
[0068] Specifically, the secondary side of the substation has several busbar sections, and the emergency power station has N busbar sections; the construction of the emergency power supply network includes:
[0069] The substation secondary side busbars are divided into N groups, each corresponding to a busbar in the emergency power station. Each busbar in each group on the secondary side of the substation is connected to the corresponding busbar in the emergency power station via an independent transmission line. Corresponding substation-side incoming circuit breakers and emergency power station-side outgoing circuit breakers are installed on the transmission lines.
[0070] This disclosure is as follows Figure 2 The illustrated embodiment will be used as an example for explanation, such as... Figure 2 As shown, the emergency power supply network includes a 110kV substation and a 10kV emergency power station, where the 10kV emergency power station is a switching station. The 110kV substation is a typical 110kV-10kV step-down substation. This means the primary side of the substation is the 110kV incoming line side, and the secondary side is the 10kV outgoing line side. The substation is configured with two 110kV dual-incoming lines in a single-branch configuration, and the 10kV substation uses a conventional 110kV substation power supply mode with four busbars in a double-branch configuration. Specifically, the two incoming lines are each connected to the 10kV busbar via a main transformer. Figure 2 The two main transformers are shown as Transformer #1 and Transformer #2, respectively. The four 10kV busbars are shown as 10kV Bus I, 10kV Bus II, 10kV Bus III, and 10kV Bus IV. Transformer #1 is connected to Bus I and Bus II via its No. 1 incoming line 901 and No. 2 incoming line 902. Transformer #2 is connected to Bus III and Bus IV via its No. 2 incoming line 903 and No. 2 incoming line 904, respectively. The IV busbars are connected; among them, the No. 1 incoming line 901 of the No. 1 main transformer, the No. 2 incoming line 902 of the No. 1 main transformer, the No. 1 incoming line 903 of the No. 2 main transformer, and the No. 2 incoming line 904 of the No. 2 main transformer are all equipped with main transformer circuit breakers with corresponding numbers (in the following description, the circuit breaker is directly referred to by its number, for example, 901 is used to represent the circuit breaker on the No. 1 incoming line 901 of the No. 1 main transformer). A bus tie circuit breaker 913 is configured between the 10kV I busbar and the 10kV III busbar, and a bus tie circuit breaker 924 is configured between the 10kV II busbar and the 10kV IV busbar.
[0071] The 10kV emergency power station has two busbars, schematically designated as 10kV Bus I and 10kV Bus II, with a bus tie circuit breaker 930 installed between them. The 10kV Bus I of the 10kV emergency power station is connected to both the 10kV Bus I and 10kV Bus II of the 110kV substation via tie lines. An emergency power station-side outgoing circuit breaker (Emergency Section 1 911) and a substation-side incoming circuit breaker (Emergency Power Supply No. 1 Incoming Line 991) are installed between the 10kV Bus I of the emergency power station and the 10kV Bus I of the 110kV substation. An emergency power station-side outgoing circuit breaker (Emergency Section 2 912) and a substation-side incoming circuit breaker (Emergency Power Supply No. 2 Incoming Line 992) are installed between the 10kV Bus I of the emergency power station and the 10kV Bus II of the 110kV substation. Similarly, the 10kV II bus of the 10kV emergency power station is connected to the 10kV III bus and 10kV IV bus of the 110kV substation via tie lines. An emergency power station-side outgoing circuit breaker (Emergency Section 3 921) and a substation-side incoming circuit breaker (Emergency Power Supply No. 3 Incoming Line 993) are installed between the 10kV II bus of the emergency power station and the 10kV III bus of the 110kV substation. An emergency power station-side outgoing circuit breaker (Emergency Section 4 922) and a substation-side incoming circuit breaker (Emergency Power Supply No. 4 Incoming Line 994) are installed between the 10kV II bus of the emergency power station and the 10kV IV bus of the 110kV substation.
[0072] Understandable Figure 2 The emergency power supply network shown is merely an illustrative representation of this disclosure, and the methods proposed in this disclosure are equally applicable to other types of emergency power supply network structures.
[0073] S2. Configure the core circuit breaker as an electrically controlled circuit breaker, wherein the electrically controlled circuit breaker performs on / off operations according to control signals; the core circuit breaker is a circuit breaker related to the automatic transfer of emergency power.
[0074] By configuring or modifying step S2, a hardware foundation is provided for converting traditional manual operation to electric operation, thereby solving the problem of long manual operation time. For newly built sites, the above configuration can be directly achieved through equipment selection. This disclosure uses the renovation of an old site as an example. For instance, a certain 110kV substation is an old site, and all 64 switchgear circuit breaker trolleys in the site are manually operated. If a fault occurs that requires the connection of emergency power, the operation steps are numerous, and the manual cranking time will increase with the number of operators and their physical strength. Therefore, the core circuit breaker trolley, such as the 10kV incoming switch ( Figure 2 The main transformer circuit breaker is located in bays 901, 902, 903, and 904, and the bus tie circuit breaker is shown. Figure 2 As shown in bays 913 and 924), emergency power supply incoming switch ( Figure 2Ten circuit breaker trolleys (intervals 991, 992, 993, and 994) were converted from manual to electric entry and exit trolleys. To save on retrofit costs, the original circuit breaker trolley chassis were replaced with electric chassis, along with a circuit breaker trolley drive and monitoring unit and an electric operation selection switch, enabling the circuit breaker trolleys to electrically rock in and out between the working and test positions. Testing verified that the time for all 10 circuit breakers to simultaneously enter and exit the trolley was reduced to 30 seconds.
[0075] Specifically, the addition of electric rocking-in / rocking-out operation function to the circuit breaker trolley is achieved by adding a motor drive mechanism to the circuit breaker chassis, enabling electric rocking-in / rocking-out operation between the running position and the isolation / test position. The motor drive mechanism is added to the circuit breaker chassis, and the motor within the drive mechanism drives the entire transmission mechanism to complete the electric rocking-in / rocking-out operation of the circuit breaker trolley. When the motor stops or fails, the clutch mechanism automatically disengages, and the movement of the spindle is disengaged from the worm gear transmission, allowing for manual rocking-in / rocking-out of the trolley at any time.
[0076] The motor-driven handcart control unit is a multi-functional controller for protecting the circuit breaker handcart electric chassis. When the drive motor is obstructed (e.g., due to improper installation or other reasons causing malfunction) and the protection condition is met, it immediately brakes the motor and reverses the drive motor to release the jammed state, making it possible to switch from electric to manual operation. It can also independently switch between motor drive and various operating states, replacing multiple self-holding relays and intermediate relays in the original control method, reducing electrical nodes, increasing system reliability, and facilitating maintenance.
[0077] The modified circuit breaker truck has the following features:
[0078] ① It has two complete locking mechanisms, one manual and one electric, which can be easily switched at any time as needed;
[0079] ② The electric operation is accurate and precise, avoiding the risk of damaging the mechanical interlocking mechanism due to incorrect human operation;
[0080] ③ By remotely operating the circuit breaker, an intelligent unmanned substation can be realized, further improving the safety and reliability of the equipment.
[0081] The typical steps for converting an electric handcart are as follows:
[0082] The circuit breaker trolley chassis was replaced with an electric trolley chassis.
[0083] 10kV switchgear indoor secondary installation of intelligent monitoring unit and wiring;
[0084] A changeover switch and wiring are added to the door panel of the secondary compartment of the 10kV switchgear; an opening needs to be made in the door panel of the secondary compartment on site;
[0085] Additional wiring was added to the secondary aviation connectors of the on-site switchgear and circuit breakers.
[0086] In some embodiments, to avoid potential safety hazards such as mechanical failures, vehicle entry / exit failures, and incomplete or loose connections of the moving and stationary contacts that might occur when the core circuit breaker is replaced with an electric trolley, visualization of the circuit breaker's moving and stationary contacts is enhanced to provide comprehensive monitoring of both indoor and indoor contacts. A night vision camera is added to the circuit breaker baffle, and the video signal is uploaded to the dispatch and monitoring computer via a monitoring unit. This facilitates real-time observation of the core switch's operational status and increases the reliability of switching operations.
[0087] S3. Collect substation operation data, which includes at least voltage data, current data, protection interlocking signals, and switch information of the collection points.
[0088] In one specific embodiment, the collected substation operation data mainly includes the following signals:
[0089] The three-phase currents of the low-voltage side incoming lines 901, 902, 903, and 904 of the main transformer in the 110kV substation;
[0090] Phase A and B voltages of 10kV bus I, II, III, and IV in a 110kV substation;
[0091] Emergency power station 10kV Bus I and Bus II three-phase voltage;
[0092] The position signals and blocking signals of circuit breakers 901, 902, 903, and 904 on the low-voltage side of the main transformer of the 110kV substation, and the position signals of circuit breakers 913 and 924 on the bus tie circuit breaker.
[0093] Handcart position signals for substations 991, 992, 993, and 994;
[0094] The position signals of the circuit breakers and the handcarts on feeders 911, 912, 921, and 922 of the emergency power station, and the position signal of the bus tie circuit breaker 930;
[0095] Other: Automatic transfer switch activation signal + power supply self-reset signal + protection reset signal.
[0096] Of course, it can also collect three-phase voltage and current data, as well as meteorological data, from the high-voltage side of the main transformer of the 110kV substation as needed.
[0097] In some embodiments, a combination of hard-wired acquisition and GOOSE soft message acquisition is used to obtain various signals. Specifically, hard-wired acquisition is used for the core bays, while GOOSE soft message acquisition is used for the remaining bays. For example, in... Figure 2 In the illustrated embodiment, all analog quantities related to the emergency power automatic transfer logic are acquired using hard-wiring. Hard-wiring acquisition is used for important bays such as the 10kV incoming switch bays (901, 902, 903, 904), bus tie bays (913, 924), emergency power incoming switch bays (991, 992, 993, 994), and emergency power station outgoing switch bays (911, 912, 921, 922) of the 110kV substation. All other bay switch signals are acquired from the relevant bay protection and control devices using GOOSE messages.
[0098] The above scheme simplifies the amount of hard wiring, reduces the wiring pressure of related automatic transfer switches, collects analog signals and important switch signals in a conservative manner, and collects other signals in a soft message manner, taking into account safety, feasibility, and cost savings.
[0099] S4. Based on the substation operation data, determine whether the current operating status of the substation requires the activation of emergency power supply.
[0100] Specifically, the aforementioned determination process is executed by the automatic transfer device according to the set switch action control logic or preset algorithm. The automatic transfer device can be configured with two devices to distribute the switching signals, with each device serving as a backup for the other. In one specific embodiment, each acquired signal is extended within the panel using relays before being connected to two automatic transfer devices, and the current loops are connected in series to both devices. Non-critical load outlets are extended via intermediate relays after passing through protection outlets. Simultaneously, a hard-plate wiring method is adopted, facilitating safe switching of the first batch of automatic transfers, with flexible and controllable intervals for emergency power restoration, further improving the device's reliability.
[0101] It should be noted that in this embodiment of the disclosure, emergency power supply is only permitted to be started when all busbar sections on the secondary side of the substation have lost voltage, or when it is determined that the fault will cause all busbar sections on the secondary side of the substation to lose voltage.
[0102] Fault types that can cause a complete loss of voltage on all busbar sections of the substation's secondary side include: loss of power to all incoming lines on the primary side of the substation, failure of all main transformers in the substation, and failure of all main transformer circuit breakers in the substation. For example, a simultaneous loss of power to the 110kV double-circuit incoming lines due to line faults, faults in upstream substations, or weather conditions; or simultaneous failures of the No. 1 and No. 2 main transformers in the 110kV substation. Different faults will cause the substation's operating data to show different states, but the final result will be a complete loss of voltage on all busbar sections of the substation's secondary side.
[0103] In some embodiments, in step S4, it is directly determined whether emergency power supply needs to be started based on the comparison result between the real-time operation data of the substation and the system preset value; or based on the historical operation data and real-time operation data of the substation, a neural network model is used to predict or detect the substation fault situation, and it is determined whether emergency power supply needs to be started based on the fault situation.
[0104] In the first judgment scheme, system preset values can be set based on typical values to identify abnormal system conditions. For example, when judging whether the 10kV I bus of a 110kV substation is energized or de-energized, the collected A and B phase voltages, i.e., line voltage U, can be used. AB Make a judgment when the line voltage U AB When the voltage exceeds the set value, the 10kV I bus is considered to be under voltage. When the line voltage U... AB When the voltage is less than the set value, the 10kV I bus is considered to be unvoltaged. The 10kV II bus, 10kV III bus and 10kV IV bus of the 110kV substation are judged based on the same criteria.
[0105] When determining whether the main transformer circuit breakers (901, 902, 903, 904) have current, based on the collected three-phase current data, if any one of the three phases has current, the main transformer circuit breaker is considered to have current; only when none of the three phases have current is the main transformer circuit breaker considered to have no current.
[0106] When determining whether the 10kV I bus and 10kV II bus of an emergency power station are energized or de-energized, the system is based on the collected three-phase voltage data and considers the system to be energized only when all three phases are energized.
[0107] The aforementioned typical value-based judgment method can effectively identify whether the current operating status of a substation requires the activation of emergency power supply. However, this method has certain shortcomings. Specifically, the transient data of the power grid varies greatly, and the power grid fluctuations differ when different fault causes lead to a complete loss of voltage on the secondary side, resulting in different states for each collected quantity. Using the above method, emergency power can only be activated after the fault occurs and stabilizes, delaying the power restoration time. This has a significant impact on the reliability of power supply, especially for important primary loads. Therefore, this disclosure designs a second judgment method that can identify fault conditions in the early stages of a fault and can also predict the occurrence of a fault in advance. This method does not rely entirely on the final result of "loss of voltage on all busbar sections on the secondary side of the substation" to determine the emergency power supply demand, thereby enabling the emergency power supply to be activated in advance, identifying faults early, predicting fault risks, reducing waiting time, shortening power outage time, and accelerating the restoration of power supply to important loads. The specific judgment method uses a neural network model including convolutional neural networks and long short-term memory networks to determine the fault type.
[0108] Specifically, the use of neural network models to predict or detect substation faults, such as... Figure 3 As shown, it includes:
[0109] S401. Obtain historical operating data of the substation arranged in chronological order and its corresponding fault information;
[0110] Historical data on switch quantities (such as the status of each circuit breaker and disconnector), voltage signals (bus voltage, voltage of each branch, etc.), and current signals (current of each branch) are acquired based on the current emergency power supply network. The acquisition frequency can be set according to actual needs, for example, once per minute.
[0111] S402. Preprocess the historical operating data of the substation, including data cleaning and normalization.
[0112] Data cleaning involves checking the integrity and accuracy of the data, removing outliers (such as voltage or current values that are significantly outside the normal range) and noisy data; filling or interpolating missing data to ensure data continuity.
[0113] Normalization involves normalizing numerical data such as voltage and current to the [0, 1] interval so that neural networks can process them more effectively. For example, for voltage data U, the normalization method is as follows:
[0114]
[0115] in, This is the normalized value of the voltage data; This is the minimum value of the voltage; This represents the maximum voltage value.
[0116] In some embodiments, the preprocessing process further includes data augmentation to increase the number of samples. For time-series data, this can be achieved through time shifting or scaling; for frequency-domain data, it can be achieved through frequency shifting or scaling; and for switching data, it can be achieved through state flipping or state sequence generation. Alternatively, the number of samples can be increased by introducing an external dataset. This external dataset should have the same substation backup power switching task and similar site structure as the current dataset, and then be merged with the original dataset. For newly added data, experts can manually assign corresponding fault condition labels.
[0117] S403. Extract features from the preprocessed historical operating data of the substation and combine them with the corresponding fault condition labels to form an operating dataset.
[0118] This step is used to extract time-domain, frequency-domain, and switching characteristics from various types of data in the historical operation data of the substation. For time-domain characteristics, statistics such as the mean, variance, peak value, and RMS value of the voltage and current signals are calculated. These characteristics reflect the steady-state and transient characteristics of the signals. For frequency-domain characteristics, Fourier transforms are performed on the voltage and current signals to extract their frequency domain features, such as the amplitude and phase of each frequency component. Fault signals often exhibit anomalies in specific frequency ranges. For switching characteristics, the switching states are converted into numerical features; for example, a circuit breaker closing is represented by a 1, and opening by a 0. Simultaneously, the rate of change of the switching states is considered, such as the number of times the switching state changes per unit time, which reflects the dynamic changes of the system.
[0119] S404. The neural network model is trained based on the running dataset; wherein, the Adam optimizer and cross-entropy loss function are used to train the neural network model to obtain the trained model.
[0120] Specifically, the running dataset obtained in step S403 is divided into a training set, a validation set, and a test set, with a ratio of 7:2:1. This ensures that the distribution of the dataset can cover various fault conditions and normal operating conditions.
[0121] The cross-entropy loss function can be well used to handle fault classification tasks.
[0122] Use the Adam optimization algorithm for training, and set an appropriate learning rate (e.g., 0.001) and number of iterations (e.g., 1000).
[0123] Evaluate the model's performance on the validation set using metrics such as accuracy, recall, and F1 score. Adjust the model structure and hyperparameters based on the evaluation results.
[0124] S405. Based on the real-time operation data of the substation, use the trained neural network model to determine the current fault type of the substation.
[0125] The real-time collected feature data is input into the trained neural network model. Based on the fault probability output by the model, when the probability of a certain fault type exceeds a set threshold (such as 0.8), the fault is determined to have occurred.
[0126] After determining the current fault type of the substation based on the neural network model, if the current fault type will cause all busbar sections on the secondary side of the substation to lose voltage, it is determined that emergency power supply needs to be activated. It is understood that the current fault type can indicate both the fault occurring at the current moment and the fault occurring at the next moment, and can be set as needed.
[0127] In some embodiments, the network architecture of the neural network model includes an input layer, a convolutional layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence.
[0128] Figure 4 A network architecture diagram of a neural network model in a specific embodiment is shown. The input layer is used to input preprocessed multidimensional time series data.
[0129] The convolutional layer uses one-dimensional convolution to extract local features from multi-dimensional time series data; here, two convolutional layers are used, with 64 and 128 filters respectively, and the kernel size is 3.
[0130] LSTM layers are used to capture long-term dependencies in time series data; specifically, two LSTM layers can be used to capture long-term dependencies in time series. The first LSTM layer returns the sequence so that the second LSTM layer can process it further.
[0131] Fully connected layers are used to integrate the extracted local features; specifically, two fully connected layers can be used to further process the extracted features, with Dropout used after each layer to prevent overfitting.
[0132] The output layer uses the softmax activation function to output the probability distribution of fault types; when the probability of a certain fault type in the output structure exceeds a set threshold, the system is considered to have encountered that fault.
[0133] In some embodiments, when the neural network model in the above-described determination method predicts that the substation fault type at the next moment will cause all busbar sections on the secondary side of the substation to lose voltage, the emergency power supply can be put into a hot standby state before starting the emergency power supply. In the hot standby state, the outgoing circuit breaker on the emergency power supply station side is in operation, and the incoming circuit breaker on the substation side is in a hot standby state. This improves the efficiency of subsequent power supply restoration. It is understood that all preparations for the activation of the emergency power supply can be performed in the hot standby state.
[0134] S5. When the current operating status of the substation indicates that emergency power supply needs to be started, the system determines whether the emergency power automatic transfer switching conditions are met based on the real-time operating data of the substation's external emergency power automatic transfer system. If the conditions are met, the system sends control signals to each core circuit breaker. The control signals control each core circuit breaker to supply emergency power to the secondary busbar of the substation according to the automatic transfer switch operation control logic.
[0135] In some embodiments, the criteria for determining whether the emergency power automatic transfer conditions are met based on real-time operating data of the substation's external emergency power automatic transfer system include:
[0136] All busbar sections on the secondary side of the substation lost power;
[0137] There is no current in the circuit where the main transformer circuit breaker is located;
[0138] The circuit breaker of the main transformer is located on a line without a protection lockout signal.
[0139] The busbar section of the emergency power station currently scheduled to be put into operation has been charged. Specifically, if the secondary busbar section of the substation is energized before the fault occurs, the outgoing circuit breaker of the emergency power station is in operation, the incoming circuit breaker of the substation is in cold standby or hot standby status, and the busbar section of the emergency power station is energized for a preset time (such as 20 seconds), it is determined that the busbar section of the emergency power station currently scheduled to be put into operation has been charged.
[0140] In some embodiments, the automatic switching action control logic includes:
[0141] Disconnect all feeder switches corresponding to the secondary busbar section of the substation.
[0142] Disconnect the main transformer circuit breaker corresponding to the secondary busbar section of the substation.
[0143] Control the closing of the substation-side incoming circuit breaker corresponding to the secondary busbar section of the substation.
[0144] After the secondary busbar section of the substation is energized, at least some of the corresponding feeder switches are closed in sequence.
[0145] In one specific embodiment, after determining that the emergency power automatic transfer procedure needs to be activated, the execution process of the emergency power automatic transfer procedure (taking the restoration of 10kV I bus power supply of a 110kV substation as an example) is as follows:
[0146] like Figure 5 As shown, after the automatic transfer switch is activated, it waits 15 seconds before execution. First, the operating mode is determined: is the 10kV I bus of the 110kV substation powered by the main transformer circuit breaker 901 or by the bus tie circuit breaker 913? If powered by the 913 bus tie switch, it is also necessary to determine whether the main transformer circuit breaker 903 has no current. The corresponding main transformer circuit breaker is then disconnected, all feeder switches are disconnected, and the corresponding substation-side incoming circuit breaker is closed. This involves retracting the 901 handcart, disconnecting all feeders, and retracting the 991 handcart.
[0147] When disconnecting all feeder switches, determine if there is an abnormal feeder position based on the feeder on / off knob. If a feeder switch knob is turned off, the switch must be in the open position before proceeding to the next step. If a feeder is not disconnected (knob off), issue a trip command again after 20 seconds until tripping is complete.
[0148] When restoring power to loads, loads can be selectively restored based on their importance and the capacity of the emergency power station. In practice, the system automatically closes the first batch of feeder switches according to the preset first batch closing logic, while the remaining backup loads are manually remotely connected based on the remaining capacity of the 10kV emergency power supply.
[0149] In some embodiments, based on the substation external emergency power supply automatic transfer system modification method proposed in this disclosure, a backup automatic transfer test design was carried out, including customized verification, backup transfer success test, no full substation voltage loss test, incoming switch current blocking test, unsuccessful circuit breaker tripping test, external blocking verification, etc.
[0150] For 110kV substations that cannot be completely shut down, especially important substations with primary loads, where a full-station shutdown for coordinated commissioning is not possible, how to commission and verify the automatic transfer system (ATS) is the biggest challenge after the ATS upgrade. To ensure the safety of operating equipment and the reliability of power supply, a four-section busbar commissioning method is adopted. The loads of sections I, II, III, and IV of the 110kV substation are emptied, and, without a complete shutdown of the main substation, switching simulations are performed on selected sections of the de-energized busbars for live-line commissioning verification. To ensure the success of this verification work, the following special commissioning method is adopted for the safety assurance of the ATS test. The special commissioning method is as follows:
[0151] First, a software program simulation is performed. The preset switching logic is tested offline and then online using the software to verify the correctness of the operation logic.
[0152] For the power outage bus section with emptied load at the main station, simulate two power outage intervals for switching and conduct live-line commissioning verification.
[0153] Extend the incoming line of the test bus and the aviation plug of the bus tie circuit breaker, and place the circuit breaker outside the switch cabinet to prevent the risk of accidental closing when the switch moving contact is inserted into the live bus side.
[0154] Disconnect all communication network cables of protection devices and switches except for the test bus section to avoid the risk of circuit breaker tripping or closing caused by erroneous logic commands.
[0155] Templated operation tickets for external emergency power supply activation and main power restoration are used to prevent misoperation caused by personnel skill level or condition.
[0156] Develop relevant emergency plans to ensure safe and reliable automatic switching of emergency power supplies, including activation conditions, one-click sequential automatic switching, and handling of automatic switching system failures.
[0157] In one specific embodiment, the automatic switching transmission of the entire group was verified for each of the four busbar segments.
[0158] The four busbar sections have the same overall transmission method, with Figure 2 Taking the 110kV substation section I busbar as an example, where 901 is the power source, some feeder circuit breakers are allowed to close. The operation process is as follows:
[0159] The automatic transfer switch is actually connected to the voltage of the 10kV I section busbar of the 110kV substation and the I and II section busbars of the 10kV emergency station. The voltage of the II, III and IV section busbars of the 110kV substation is brought to normal by the relay protection device. After the device is charged, the voltage circuit breaker of the 10kV I section PT (voltage transformer) is disconnected at the 10kV I section PT (voltage transformer) cabinet to simulate the loss of voltage of the 10kV I section busbar. The voltage of the 10kV II, III and IV section busbars of the 110kV substation is reduced to simulate the loss of voltage of the 10kV II, III and IV sections.
[0160] Undervoltage start-up automatic transfer: ① After 15 seconds, the device trips 901 and the load feeder circuit breaker, i.e., the feeder switch; ② The 901 circuit breaker truck is withdrawn, the 913 circuit breaker truck is withdrawn, and the 991 circuit breaker truck is brought in; ③ The 991 circuit breaker is closed; ④ After the 911 circuit breaker is closed, the 10kV I section busbar is simulated to be restored; ⑤ The automatic transfer device closes some important feeder switches, while the remaining feeder switches are not closed; ⑥ After the automatic transfer reaches the set delay, the device reports successful transfer, and the process ends.
[0161] Test conclusion: The device operates correctly, and all power supply and feeder circuit breakers open and close correctly according to the preset logic.
[0162] In this specification, the illustrative expressions of the terms used do not necessarily refer to the same embodiments or examples. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0163] Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this invention shall be included within the scope of protection of this invention.
Claims
1. A method for retrofitting an external emergency power supply automatic transfer system for a substation, characterized in that, include: Construct an emergency power supply network; wherein, the substation includes a primary bus and a secondary bus, and a power supply link is constructed between the emergency power station and the secondary bus of the substation, and the power supply voltage level of the emergency power station and the secondary bus of the substation is the same. The core circuit breaker is configured as an electrically operated circuit breaker, which performs on / off operations according to control signals; the core circuit breaker is a circuit breaker related to the automatic transfer of emergency power. Collect substation operation data, which includes at least voltage data, current data, protection interlocking signals, and switch information at the collection points; Based on the substation's operating data, determine whether the current operating status of the substation requires the activation of emergency power supply; When the current operating status of the substation indicates that emergency power supply needs to be started, the system determines whether the emergency power automatic transfer switching conditions are met based on the real-time operating data of the substation's external emergency power automatic transfer system. If the conditions are met, the system sends control signals to each core circuit breaker. The control signals then control each core circuit breaker to supply emergency power to the substation's secondary bus according to the automatic transfer switch operation control logic. The step of determining whether the current operating status of the substation requires the activation of emergency power supply based on substation operating data includes: Based on historical and real-time operating data of the substation, a neural network model is used to predict or detect substation faults, and to determine whether emergency power supply needs to be activated based on the fault situation. The neural network model includes convolutional neural networks and long short-term memory networks; The method of using a neural network model to predict or detect substation faults includes: Obtain historical operating data of substations arranged in chronological order and their corresponding fault information; Preprocessing of historical operating data of substations includes data cleaning and normalization. Feature extraction is performed on the preprocessed historical operation data of the substation, and combined with the corresponding fault condition labels to form an operation dataset; The neural network model is trained based on the running dataset; wherein, the Adam optimizer and cross-entropy loss function are used to train the neural network model to obtain the trained model; Based on real-time operating data of the substation, the current fault type of the substation is determined using a trained neural network model. After determining the current fault type of the substation based on the neural network model, if the current fault type of the substation will cause all busbar sections on the secondary side of the substation to lose voltage, it is determined that emergency power supply needs to be started.
2. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 1, characterized in that, The substation secondary side has several busbar sections, and the emergency power station has N busbar sections; the construction of the emergency power supply network includes: The substation secondary side busbars are divided into N groups, each corresponding to a busbar in the emergency power station. Each busbar in each group on the secondary side of the substation is connected to the corresponding busbar in the emergency power station via an independent transmission line. Corresponding substation-side incoming circuit breakers and emergency power station-side outgoing circuit breakers are installed on the transmission lines.
3. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 2, characterized in that, At least one main transformer is installed between the primary busbar and the secondary busbar of the substation, and the secondary side of each main transformer is connected to the corresponding secondary busbar section through a main transformer circuit breaker. Emergency power supply is only permitted to be activated when all busbar sections on the secondary side of the substation are de-energized. Fault types that can cause all busbar sections on the secondary side of a substation to lose power include: loss of power to all incoming lines on the primary side of the substation, failure of all main transformers in the substation, and failure of all main transformer circuit breakers in the substation.
4. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 3, characterized in that, The network architecture of the neural network model includes: The input layer receives preprocessed multidimensional time series data. Convolutional layers use one-dimensional convolution to extract local features from multi-dimensional time series data; LSTM layers are used to capture long-term dependencies in time series data; Fully connected layers are used to integrate extracted local features; The output layer uses the softmax activation function to output the probability distribution of fault types; when the probability of a certain fault type in the output structure exceeds a set threshold, the system is considered to have encountered that fault.
5. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 4, characterized in that, When the neural network model predicts that the substation fault type will cause all busbar sections on the secondary side of the substation to lose voltage in the next moment, the emergency power supply will be put into hot standby state before the emergency power supply is started. In the hot standby state, the outgoing circuit breaker on the emergency power station side is in operation, and the incoming circuit breaker on the substation side is in hot standby state.
6. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 1, characterized in that, The criteria for determining whether the emergency power automatic transfer conditions are met based on real-time operating data of the substation's external emergency power automatic transfer system include: All busbar sections on the secondary side of the substation lost power; There is no current in the circuit where the main transformer circuit breaker is located; The circuit breaker of the main transformer is located on a line without a protection lockout signal. The busbar section of the emergency power station currently scheduled to be put into operation has been charged. Specifically, if the secondary busbar section of the substation was energized before the fault occurred, the outgoing circuit breaker of the emergency power station was in operation, the incoming circuit breaker of the substation was in cold standby or hot standby status, and the busbar section of the emergency power station was energized for a preset period of time, it is determined that the busbar section of the emergency power station currently scheduled to be put into operation has been charged.
7. The method for upgrading a substation external emergency power supply automatic transfer system according to claim 1, characterized in that, The automatic switching action control logic includes: Disconnect all feeder switches corresponding to the secondary busbar section of the substation. Disconnect the main transformer circuit breaker corresponding to the secondary busbar section of the substation. Control the closing of the substation-side incoming circuit breaker corresponding to the secondary busbar section of the substation. After the secondary busbar section of the substation is energized, at least some of the corresponding feeder switches are closed in sequence.
Citation Information
Patent Citations
Method and system for optimizing action logic of substation incoming line spare power automatic switching device
CN116914907A