A power distribution station house comprehensive online state monitoring device

Abstract

The application relates to the technical field of power distribution station monitoring, and particularly discloses a comprehensive online state monitoring device for a power distribution station house, which comprises a multi-element data acquisition module, which is used for acquiring charging state information and operation state information of a backup automatic switching device, as well as station house power consumption parameters and equipment power consumption parameters; a state analysis and prediction module, which is connected with the multi-element data acquisition module, is used for receiving the collected data, performing real-time analysis on the state of the backup automatic switching device, simultaneously performing abnormal prediction and early warning based on the station house power consumption parameters and the equipment power consumption parameters, and judging whether potential faults and hidden dangers exist; a remote control module, which is connected with the state analysis and prediction module, is used for receiving remote instructions, performing remote starting or stopping control on the backup automatic switching device according to the remote instructions; and a data transmission and management module. The technical scheme of the application can comprehensively detect, save and analyze data, and realize early warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution station monitoring, in particular to a comprehensive online state monitoring device for power distribution station. BACKGROUND

[0002] In modern power systems, power distribution stations are critical nodes for power distribution, and their stability and safety are of great importance. With the rapid development of social economy, the demand for electricity continues to grow, and higher requirements are placed on the monitoring and management of power distribution stations.

[0003] At present, although some progress has been made in the monitoring technology for power distribution stations, there are still many deficiencies. Some monitoring systems can only monitor a single electrical parameter or environmental parameter, and cannot fully reflect the overall operation state of the power distribution station. For example, traditional monitoring methods cannot simultaneously detect the comprehensive detection of backup automatic switching devices, station house alternating current power, and equipment direct current power, which leads to the inability to timely discover potential fault hazards.

[0004] In terms of detection of backup automatic switching devices, existing technologies cannot accurately obtain key information such as charging state and operation mode, and it is also difficult to realize real-time monitoring and remote control of the state of the backup automatic switching device. This makes it difficult for the backup automatic switching device to act in a timely and accurate manner when the power system is abnormal, affecting the reliability of power supply.

[0005] For the detection of station power (alternating current) and equipment power (direct current), existing monitoring equipment often has single function, only monitoring parameters such as voltage and current, although it can issue an alarm in time for overvoltage, undervoltage and other abnormal conditions, but it cannot make early prediction and warning according to relevant data, and cannot exclude faults in advance.

[0006] In addition, in terms of data transmission and management, existing technologies cannot fully utilize the existing DTU equipment in the station to efficiently transmit monitoring data to the monitoring room, and the recording and saving of data is not complete and standardized, which brings great difficulty to subsequent fault analysis and operation and maintenance management.

[0007] In summary, the existing monitoring technology for power distribution stations cannot meet the needs of comprehensive detection, data saving and analysis, and early warning. SUMMARY

[0008] The present application provides a comprehensive online state monitoring device for power distribution station, which can perform comprehensive detection, data saving and analysis, and early warning.

[0009] To solve the above technical problems, the present application provides the following technical solutions:

[0010] A comprehensive online state monitoring device for power distribution station, comprising:

[0011] A multi-element data acquisition module is configured to acquire the charging state information and the running state information of the backup automatic switching device, and the power consumption parameters of the station building and the equipment;

[0012] A state analysis and prediction module is connected with the multi-element data acquisition module, configured to receive the acquired data, and perform real-time analysis on the state of the backup automatic switching device, and perform abnormal prediction and early warning based on the power consumption parameters of the station building and the equipment, to determine whether there is a potential fault hidden danger;

[0013] A remote control module is connected with the state analysis and prediction module, configured to receive a remote instruction, and perform remote enabling or disabling control on the backup automatic switching device according to the remote instruction;

[0014] A data transmission and management module is connected with the multi-element data acquisition module, the state analysis and prediction module, and the remote control module, configured to transmit the acquired data and the analysis result to the monitoring room by using the existing DTU device of the station building, and record and save the data completely and regularly, and further configured to send early warning information to the monitoring room through the data transmission and management module when the state analysis and prediction module determines that there is an abnormality.

[0015] Further, the multi-element data acquisition module comprises:

[0016] A backup automatic switching state acquisition submodule is configured to acquire the charging state information and the running state information of the backup automatic switching device; the backup automatic switching state acquisition submodule establishes a data connection through a communication interface with the backup automatic switching device, reads the charging flag bit and the switch position signal data stored in the backup automatic switching device according to a preset communication protocol; a character “C” is used to represent the charging state, and “C” is assigned a value of “1” when the charging is completed, and a value of “0” when the charging is not performed; a character “O” is used to represent the running state, and “O” is assigned a value of “1” when the running is normal, and a value of “0” when the running is abnormal; then, the acquired data is subjected to CRC check, and if the check fails, the data is reacquired;

[0017] A station building power consumption acquisition submodule is configured to acquire the power consumption parameters of the station building; the station building power consumption acquisition submodule converts the high voltage and large current of the station building alternating current power consumption into a small signal by using a high-precision voltage transformer and a current transformer; a character “VAC” is used to represent the alternating current voltage, and the unit is volt; a character “IAC” is used to represent the alternating current, and the unit is ampere; during the acquisition process, the signal is sampled multiple times and the noise interference is removed by a digital filtering algorithm; then, a time stamp is attached to the acquired data;

[0018] The device power acquisition sub-module is used for acquiring device power parameters; the device power acquisition sub-module measures direct current by using a Hall sensor and acquires direct voltage through a voltage dividing circuit; the direct voltage is represented by a character "VDC" and the unit is volt; the direct current is represented by a character "IDC" and the unit is ampere; the sensor and the acquisition circuit are electromagnetically shielded on hardware, and then the acquisition data is checked and corrected; after the data acquisition is completed, the data is normalized, and the data is mapped to the interval [0, 1].

[0019] Further, the state analysis and prediction module performs online real-time analysis on the received data based on time series, sets a fixed time interval Δt, and in each time interval, the multi-element data acquisition module collects:

[0020] The charging state information and the running state information of the backup power automatic throw-in device are represented by characters C and O respectively;

[0021] The station building power parameters include alternating voltage V AC and alternating current I AC ;

[0022] The device power parameters include direct voltage V DC and direct current I DC are integrated and processed.

[0023] Further, for the backup power automatic throw-in device, if the charging state C changes from "0" to "1" in the continuous n time intervals, and the running state O is always "1", it is determined that the backup power automatic throw-in device is in the normal charging and ready state; if C is continuously "0" and O is "0", it is marked that the backup power automatic throw-in device may have a fault, and the fault time sequence T fault is recorded.

[0024] Further, a long short-term memory network model based on machine learning is used for abnormal prediction and early warning; the collected various data are sequentially composed into a feature vector X t =[C t , O t , V AC，t , I AC，t , V DC，t , I DC，t ], wherein t represents a time step;

[0025] The core formula of the long short-term memory network model based on machine learning is:

[0026] The forgetting gate f t =σ(W f ·[h t-1 , X t ]+b f), which is used to determine which information at the last time step needs to be preserved to the current time step, and σ is a Sigmoid function, W f is the weight matrix of the forget gate, h t-1 is the hidden state at the last time step, b f is the bias vector of the forget gate.

[0027] The input gate i t = σ(W i · [h t-1 , X t ]+b i ), is used to determine how to add new information at the current time step to the cell state, W i , W c are the corresponding weight matrices, b i , b c are the bias vectors, and tanh is the hyperbolic tangent function.

[0028] The cell state updates the cell state.

[0029] The output gate o t = σ(W o · [h t-1 , X t ]+b o ), h t = o t *tanh(C t ), which calculates the hidden state and output at the current time step.

[0030] Further, in the model training phase, a large amount of historical data is used for training, and the historical data includes normal operation data and various fault data; a cross-validation method is used to divide the data set into k non-overlapping subsets, each time using \(k-1\) subsets for training and the remaining one subset for validation, repeating k times, and taking the average validation result as the model performance evaluation index.

[0031] Further, in the model training process, the hyperparameters of the long short-term memory network model based on machine learning are continuously adjusted to minimize the loss function of the model on the validation set, and the loss function adopts the mean square error; the formula is where N is the number of samples, y i is the true value, is the predicted value; when the confidence of the prediction result is lower than the set threshold α, the abnormal situation is confirmed through manual intervention and further data analysis.

[0032] Further, the state analysis and prediction module is also used to analyze the correlation between data, and a data correlation degree calculation formula is introduced to determine the rationality of data; let the data correlation degree R, the correlation degree calculation formula of the charging state C of the backup power transfer device and the alternating voltage V AC is:

[0033]

[0034] Wherein, n is the sample number of collected data, C i and V ACi : the charging state and alternating voltage data collected for the ith time, and are the average values of the charging state and alternating voltage data respectively; under normal circumstances, should be within a reasonable interval [r min , r max ];

[0035] If the calculated is out of the interval, and the number of times that the alternating voltage V AC is out of the normal range reaches a certain threshold m, it is suspected that the alternating voltage data may be caused by sensor error.

[0036] When it is suspected that a certain data is not reasonable due to sensor error, the data of other related sensors is compared; if all other data is within the normal range and differs greatly from the current abnormal data, it is determined that the current collected data is unreasonable data, the backup power transfer device state abnormal alarm caused by this data is false alarm, and no alarm signal is sent, and the abnormal data situation is recorded.

[0037] Further, the data transmission and management module will verify the analysis result before transmitting it to the monitoring room; for the backup power transfer device state analysis result, the actual state of the backup power transfer device and the analysis result will be checked again; if the analysis result shows that the backup power transfer device should be started, but the actual switch position shows that it is in the off state and not in the charging state, the analysis result is marked as suspicious; the data transmission and management module and the data transmission and management modules of other devices in the station establish a data interaction link, when receiving relevant data sent by other devices, it will make comprehensive judgment with the data collected and analyzed by itself, when the device power consumption is abnormal analyzed by the device, and similar abnormal data sent by the data transmission and management module of the adjacent device is received, the credibility of the abnormal analysis result will be improved; if the adjacent device data shows normal, the device will check the accuracy of the collected data again, and recalculate the analysis result.

[0038] Further, the data transmission and management module is also used for filtering the collected data, adopts a sliding average filtering algorithm, sets a time window T, and performs a sliding average calculation on the station power consumption parameters and the equipment power consumption parameters within the time window T.

[0039] The basic scheme principle and beneficial effects are as follows: The multi-element data collection module comprehensively collects the charging state information and the running state information of the backup power automatic throw-in device, and the station power consumption parameters and the equipment power consumption parameters through specific sensors and communication modes. These data are transmitted to the state analysis and prediction module in real time. The module uses specific algorithms and models to perform real-time analysis on the state of the backup power automatic throw-in device, and performs abnormal prediction and early warning according to the station power consumption parameters and the equipment power consumption parameters.

[0040] In the analysis process, unreasonable data caused by sensor errors and other reasons are identified through data correlation judgment and other logic to ensure the accuracy of the analysis results. The remote control module receives external instructions, and according to the results of the state analysis and prediction module, realizes remote start or stop control of the backup power automatic throw-in device. The data transmission and management module uses the existing DTU equipment of the station to transmit the collected data and analysis results to the monitoring room, and performs complete and standardized recording and saving. In the data transmission process, the analysis results are verified, and comprehensive judgment is made in combination with the data of other equipment to effectively avoid false alarms caused by power grid fluctuations and other factors.

[0041] The multi-element data collection module can simultaneously collect various key information of the backup power automatic throw-in device, station power consumption and equipment power consumption, changes the limitation of traditional monitoring methods that can only monitor single electrical parameters or environmental parameters, realizes comprehensive detection of the distribution station, can timely find potential fault hidden dangers, and improves the comprehensiveness and accuracy of the monitoring.

[0042] The data transmission and management module performs complete and standardized recording and saving of the collected data, providing a rich data basis for subsequent data analysis. The state analysis and prediction module builds a prediction model based on these data, such as a long short-term memory (LSTM) model, which can predict abnormal situations in advance. Compared with traditional monitoring equipment that can only issue an alarm when an abnormality occurs, the device realizes early warning, allowing the maintenance personnel to have more time to take measures, reducing the possibility and impact of failure.

[0043] Through real-time monitoring and analysis of the charging state and running state of the backup power automatic throw-in device, and the remote control function, it is ensured that the backup power automatic throw-in device can act in time and accurately when the power system appears abnormal. Avoids the problem that the backup power automatic throw-in device cannot respond in time due to the difficulty in accurately obtaining key information and remote control of the backup power automatic throw-in device, and improves the reliability of power supply.

[0044] The accurate monitoring and analysis of the power consumption and equipment power consumption parameters of the station house can timely find abnormal changes in voltage, current and other parameters. Through the prediction and early warning function, measures are taken in advance to adjust the power distribution, avoid equipment damage or power system failure caused by abnormal power consumption, and enhance the stability of the entire power distribution station house system operation.

[0045] The complete record and standard storage of data, as well as the effective transmission of analysis results, provide detailed and accurate data support for operation and maintenance personnel. It is convenient for operation and maintenance personnel to analyze faults and summarize experience, which helps to develop more reasonable operation and maintenance plans, improve the efficiency and quality of operation and maintenance management, and reduce unnecessary operation and maintenance costs and time waste. Through reasonable judgment of the correlation of collected data, verification of analysis results, and comprehensive judgment combined with other device data, false alarms caused by sensor errors, power grid fluctuations and other factors are effectively avoided. The alarm information received by the monitoring room is more accurate and reliable, reducing the invalid troubleshooting work of operation and maintenance personnel due to false alarms, and improving the work efficiency.

[0046] In summary, the present application realizes comprehensive detection, data storage and analysis, and early warning. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 A logic block diagram of an embodiment of a power distribution station house comprehensive online state monitoring device. DETAILED DESCRIPTION

[0048] The following will be further described in detail through specific embodiments:

[0049] The embodiment is basically as shown in Figure 1 :

[0050] A power distribution station house comprehensive online state monitoring device, comprising:

[0051] A multi-element data acquisition module for acquiring charging state information and operating state information of the backup automatic switching device, as well as station power consumption parameters and equipment power consumption parameters;

[0052] A state analysis and prediction module connected with the multi-element data acquisition module, for receiving the collected data and performing real-time analysis on the state of the backup automatic switching device, and simultaneously performing abnormal prediction and early warning based on the station power consumption parameters and equipment power consumption parameters to determine whether there is a potential fault hidden danger;

[0053] A remote control module connected with the state analysis and prediction module, for receiving remote instructions and remotely enabling or disabling the backup automatic switching device according to the remote instructions;

[0054] The data transmission and management module is connected with the multi-element data acquisition module, the state analysis and prediction module, and the remote control module, and is configured to transmit the collected data and analysis results to the monitoring room by using the existing DTU equipment in the station room, and record and save the data completely and regularly. The data transmission and management module is also configured to send a warning information to the monitoring room through the data transmission and management module when the state analysis and prediction module determines that there is an abnormality.

[0055] In specific use, the multi-element data acquisition module, the state analysis and prediction module, the remote control module and the data transmission and management module are installed at appropriate positions in the power distribution station. These modules can be integrated in a control cabinet, or can be installed according to the actual situation, but the electrical connection between the modules should be stable and the signal transmission should be reliable.

[0056] The backup power automatic throw-in device data acquisition is connected with the backup power automatic throw-in device through a dedicated communication interface (such as RS-485, Ethernet, etc.), and reads the charging state information (such as charging flag bit, charging progress, etc.) and the running state information (such as switch position, fault alarm signal, etc.) of the backup power automatic throw-in device according to the communication protocol of the backup power automatic throw-in device.

[0057] Voltage transformers, current transformers and other sensors are installed at key positions such as incoming switch cabinets and outgoing switch cabinets in the station room to collect power consumption parameters such as AC voltage, AC current, active power and reactive power. For important equipment in the power distribution station (such as transformers and capacitors), corresponding sensors are installed to collect power consumption parameters such as DC voltage, DC current, temperature and humidity.

[0058] A high-performance industrial computer or server is used as the hardware platform of the state analysis and prediction module, and corresponding data analysis software is installed. The remote control module uses programmable logic controllers (PLC) or intelligent relays, which are connected with the state analysis and prediction module through a communication interface. At the same time, the remote control module establishes a remote communication connection with the monitoring room through the network to receive remote instructions from the monitoring room. The data transmission and management module can use an embedded system or an industrial data acquisition instrument, which transmits the collected data and analysis results to the monitoring room through the existing DTU equipment (such as GPRS DTU and Ethernet DTU) in the station room. At the same time, the module is equipped with a large-capacity storage device (such as a hard disk or a solid-state disk) for complete and regular recording and saving of data.

[0059] When the system starts, each module performs initialization operations in turn. The multi-element data acquisition module checks the connection status of the sensors and the communication interface, and initializes the data acquisition parameters; the state analysis and prediction module loads the data analysis algorithm and the prediction and warning model; the remote control module checks the communication connection with the monitoring room; and the data transmission and management module checks the working status of the DTU equipment and the available space of the storage device.

[0060] After normal operation, the multi-element data acquisition module collects data at a set time interval and transmits the data to the state analysis and prediction module. The state analysis and prediction module analyzes and predicts the data in real time to determine whether there is a potential fault hazard. The operator in the monitoring room can send remote instructions to the remote control module through the remote control terminal. After receiving the instructions, the remote control module analyzes and executes the corresponding control operation. The data transmission and management module transmits the collected data and analysis results to the monitoring room through the DTU device and stores the data in the local storage device. At the same time, the early warning information is processed in a timely manner to ensure that the monitoring room can timely understand the operation state of the power distribution station house.

[0061] The multi-element data acquisition module comprises:

[0062] The backup power automatic switching state acquisition submodule is used for acquiring the charging state information and the running state information of the backup power automatic switching device. The backup power automatic switching state acquisition submodule establishes a data connection through a communication interface with the backup power automatic switching device, reads the charging flag bit and the switch position signal data stored in the backup power automatic switching device according to a preset communication protocol. The character “C” is used to represent the charging state. When the charging is completed, “C” is assigned a value of “1”, and when the charging is not completed, “C” is assigned a value of “0”. The character “0” is used to represent the running state. When the running is normal, “0” is assigned a value of “1”, and when the running is abnormal, “0” is assigned a value of “0”. Then, the collected data is subjected to CRC check. If the check fails, the data is re-collected.

[0063] The station house power consumption acquisition submodule is used for acquiring the station house power consumption parameters. The station house power consumption acquisition submodule uses high-precision voltage transformers and current transformers to convert the high voltage and large current of the station house alternating current power consumption into small signals. The character “VAC” is used to represent the alternating current voltage, and the unit is volt. The character “IAC” is used to represent the alternating current, and the unit is ampere. During the acquisition process, the signals are sampled multiple times and the noise interference is removed through a digital filtering algorithm. Then, a time stamp is attached to the acquired data.

[0064] The equipment power consumption acquisition submodule is used for acquiring the equipment power consumption parameters. The equipment power consumption acquisition submodule uses a Hall sensor to measure the direct current, and uses a voltage dividing circuit to acquire the direct current voltage. The character “VDC” is used to represent the direct current voltage, and the unit is volt. The character “IDC” is used to represent the direct current, and the unit is ampere. The sensor and the acquisition circuit are electromagnetically shielded on the hardware, and then the acquired data is checked and corrected. After the data acquisition is completed, the data is normalized, and the data is mapped to the interval [0, 1].

[0065] The state analysis and prediction module performs online real-time analysis on the received data based on time series. A fixed time interval Δt is set. In each time interval, the multi-element data acquisition module collects:

[0066] The charging state information and the running state information of the backup power automatic throw-in device are respectively represented by characters C and O;

[0067] The power consumption parameters of the station building include alternating voltage V AC and alternating current I AC ;

[0068] The power consumption parameters of the equipment include direct current voltage V DC and direct current I DC .

[0069] For the backup power automatic throw-in device, if the charging state C changes from “0” to “1” in the continuous n time intervals, and the running state O is always “1”, it is determined that the backup power automatic throw-in device is in the normal charging and ready state; if C is continuously “0” and O is “0”, it is marked that the backup power automatic throw-in device may have a fault, and the fault time sequence T fault is recorded.

[0070] The long short-term memory network model based on machine learning is used for abnormal prediction and early warning; the collected various data are sequentially composed into a feature vector X t =[C t , O t , V AC，t , I AC，t , V DC，t , I DC，t ], wherein t represents a time step;

[0071] The core formula of the long short-term memory network model based on machine learning is:

[0072] The forgetting gate f t =σ(W f ·[h t-1 , X t ]+b f ) is used to determine which information of the previous moment needs to be retained to the current moment, σ is a Sigmoid function, W f is a weight matrix of the forgetting gate, h t-1 is a hidden state of the previous moment, and b f is a bias vector of the forgetting gate;

[0073] The input gate i t =σ(W i ·[h t-1 , X t ]+b i ), is used to determine how new information of the current moment is added to the cell state, W i , W c are corresponding weight matrices, and b i , b cis a bias vector, and tanh is the hyperbolic tangent function;

[0074] Cell state Update the cell state;

[0075] Output gate o t = sigma(W o * [h t-1 , X t ] + b o ), h t = o t * tanh(C t ), the hidden state and output of the current time are calculated.

[0076] In the model training phase, a large amount of historical data is used for training, and the historical data contains normal operation data and various fault data; the cross-validation method is adopted, the data set is divided into k non-overlapping subsets, each time k-1 subsets are used for training, and the remaining one subset is used for verification, repeated k times, and the average verification result is taken as the model performance evaluation index.

[0077] In the model training process, the hyperparameters of the long short-term memory network model based on machine learning are continuously adjusted to minimize the loss function of the model on the validation set, and the loss function adopts the mean square error; the formula is Where N is the number of samples, yi is the true value, is the predicted value; when the confidence of the prediction result is lower than the set threshold alpha, the abnormal situation is confirmed through manual intervention and further data analysis.

[0078] The state analysis and prediction module is also used to analyze the correlation between data, and a data correlation degree calculation formula is introduced to judge the rationality of the data; let the data correlation degree R, for the correlation degree calculation formula of the charging state C of the backup power automatic transfer device and the alternating voltage V AC :

[0079]

[0080] Where n is the number of samples of the collected data, C i and V ACi : are the charging state and alternating voltage data collected for the i-th time, and are the average values of the charging state and alternating voltage data respectively; under normal circumstances, should be within a reasonable interval [r min , r max ];

[0081] If the calculated is outside the interval, and the alternating voltage V ACIf the number of times the error exceeds a certain threshold m, it is suspected that the AC voltage data may be unreasonable due to sensor error.

[0082] When it is suspected that a certain data is unreasonable due to sensor error, it is compared with the data of other related sensors. If the other data are all within the normal range and the difference from the current abnormal data is large, then the currently collected data is determined to be unreasonable data. The alarm for abnormal status of the backup automatic transfer device caused by this data is a false alarm, and no alarm signal is issued. The abnormal data is recorded.

[0083] Before transmitting the analysis results to the monitoring room, the data transmission and management module verifies the results. For the analysis results of the automatic transfer switch (ATS) status, it checks again whether the actual status of the ATS matches the analysis results. If the analysis results indicate that the ATS should be activated, but the actual switch position shows it is off and not charging, the analysis result is marked as suspicious. The data transmission and management module establishes a data interaction link with the data transmission and management modules of other devices within the station. When it receives relevant data from other devices, it comprehensively judges it with its own collected and analyzed data. When the device analyzes an abnormal power consumption of a device and simultaneously receives similar abnormal data from the data transmission and management modules of adjacent devices, it increases the credibility of the abnormal analysis result. If the data displayed by adjacent devices is normal, the device checks the accuracy of its own collected data again and recalculates the analysis results.

[0084] The data transmission and management module is also used to filter the collected data. It adopts a moving average filtering algorithm, sets a time window T, and calculates the moving average of the station power consumption parameters and equipment power consumption parameters within the time window T.

[0085] In practical use: Assume the automatic transfer switch (ATS) installed in the substation supports the Modbus RTU communication protocol, and the ATS status acquisition submodule is connected to the ATS via an RS485 interface. During system initialization, the system reads the charging flag and switch position signal data stored internally in the ATS according to the register address specified by the Modbus RTU protocol. For example, data is acquired every 10 seconds. When the acquired register value corresponding to the charging flag meets the charging completion condition, the character "C" is assigned a value of "1"; otherwise, it is assigned a value of "0". For the operating status, the system judges based on the switch position signal and the fault feedback information inside the device. "O" is assigned a value of "1" when normal and a value of "0" when abnormal. In one acquisition process, if the acquired data is [C=0, O=1], and the CRC check fails, the system immediately re-acquires the data until the check succeeds, ensuring data accuracy.

[0086] High-precision voltage and current transformers are installed at the incoming switchgear of the station building. For example, the voltage transformer has a ratio of 10,000:100, and the current transformer has a ratio of 200:5. When the station building uses 10 kV AC power and the current is 100 A, the voltage transformer outputs a small signal of 100 V, and the current transformer outputs a small signal of 2.5 A. The power consumption acquisition sub-module samples these signals at a frequency of 10 times per second, and then uses a median filter algorithm to remove noise interference. The collected AC voltage data "VAC" and AC current data "IAC" are attached with a time stamp accurate to milliseconds, such as "2024-12-01 10:00:00.123". At a certain time, a series of AC voltage data is collected, and after filtering, a stable "VAC = 10.1 kV" is obtained and stored with a time stamp.

[0087] For key DC devices in the distribution station building, such as backup charging power, a Hall sensor is used to measure the DC current, and a voltage dividing circuit is used to collect the DC voltage. Both the Hall sensor and the voltage dividing circuit are subjected to electromagnetic shielding treatment to reduce external electromagnetic interference. The collected data, such as "VDC = 225 V" and "IDC = 5 A", are first subjected to checksum and error correction processing, and then normalized using the formula "normalized value = (original value - minimum value) / (maximum value - minimum value)". Assuming that the minimum value of the DC voltage is 200 V and the maximum value is 250 V, then the normalized value of "VDC" is (225-200) / (250-200) = 0.5.

[0088] Set the time interval Δt = 1 minute, and integrate the collected data in each time interval. For example, in the next 5 time intervals, the charging state C of the backup device changes from "0" to "1", and the running state 0 is always "1", then it is determined that the backup device is in a normal charging and ready state. If C is "0" and 0 is "0" for a period of time, such as 10 consecutive time intervals, it is marked that the backup device may have a fault, and the fault time sequence is recorded.

[0089] The long short-term memory (LSTM) model is used for anomaly prediction and early warning. The operation data of the distribution station house in the past year are collected as historical data, including various data during normal operation and fault data such as AC voltage fluctuation and device overload. The data set is divided into 5 non-overlapping subsets (k = 5) for 5-fold cross-validation. During training, the number of hidden layer neurons and learning rate of the LSTM model are adjusted. Assuming that the initial number of hidden layer neurons is 64 and the learning rate is 0.01, after multiple tests, it is found that when the number of hidden layer neurons is adjusted to 128 and the learning rate is adjusted to 0.001, the mean square error loss function of the model on the validation set is the smallest. In a prediction, the confidence of the prediction result is 0.7, which is lower than the set threshold 0.8, at which time the system prompts manual intervention, and the operation and maintenance personnel make further analysis and confirmation combined with more field data and experience.

[0090] For data correlation analysis, it is assumed that 100 sample data are collected in a period of time, and the correlation degree of the charging state C of the backup power automatic transfer device and the AC voltage VAC is calculated If the calculated value is outside the normal interval [0.5, 0.7], and the number of times that the AC voltage VAC exceeds the normal range (for example, the normal range is 9.5-10.5 kV, V AC is less than 9.5 kV reaches the threshold value m = 10 times, it is suspected that the AC voltage data may have a problem. At this time, the AC voltage data of other monitoring points in the station are compared, and if the data of other monitoring points are within the normal range and differ greatly from the current abnormal data, such as the voltage of other monitoring points is between 10.2-10.3 kV, and the current data is 9.2 kV, it is determined that the collected AC voltage data is unreasonable, to avoid false positives due to the data, and the abnormal situation is recorded.

[0091] Before transmitting the analysis result to the monitoring room, the backup power automatic transfer device state analysis result is verified. For example, the analysis result shows that the backup power automatic transfer device should be started, but by querying the actual switch position register and charging flag of the backup power automatic transfer device, it is found that the switch is in the off state and the charging state is not charged, so the analysis result is marked as suspicious.

[0092] The data transmission and management module establishes a data interaction link with the data transmission and management modules of other devices in the station through Ethernet. When the device analyzes that the power consumption of a certain device is abnormal, such as "VDC" exceeding the normal range, and receives similar abnormal data sent by the data transmission and management module of the adjacent device, the system will increase the credibility of the abnormal analysis result from 0.6 to 0.8. If the adjacent device data shows normal, the device will recheck its own acquisition circuit and sensor, reacquire data and calculate the analysis result.

[0093] For the filtering processing of the collected data, a time window T=5 minutes is set, and the sliding average calculation is performed on the power consumption parameters of the station building and the equipment. Taking the alternating current "IAC" as an example, the data is collected every 10 seconds within 5 minutes, a total of 30 data points, the sliding average filtering algorithm is used to calculate the average value, and more stable current data is obtained for subsequent analysis and transmission, which effectively avoids the data misjudgment caused by instantaneous fluctuation.

[0094] The above is only an embodiment of the application, and the application is not limited to this embodiment. The specific structure and characteristics of the scheme known in the art are not described in detail, and the ordinary skilled person in the art knows all the ordinary technical knowledge in the technical field of the application before the filing date or the priority date, can know all the prior art in the field, and has the ability to apply conventional experimental means before that date. The ordinary skilled person in the art can improve and implement the scheme under the guidance of this application, and some typical known structures or known methods should not be an obstacle for the ordinary skilled person in the art to implement the application. It should be pointed out that, for those skilled in the art, without departing from the structure of the application, a number of modifications and improvements can be made, which should also be considered as the protection scope of the application, and these will not affect the effect and practicality of the application. The protection scope of the present application should be subject to the content of its claims, and the specific implementation mode and the like in the specification can be used to explain the content of the claims.

Claims

1. A comprehensive online status monitoring device for a power distribution station, characterized in that, include: The multi-data acquisition module is used to collect charging status information and operating status information of the automatic transfer switch, as well as power consumption parameters of the station and equipment. The status analysis and prediction module is connected to the multi-source data acquisition module. It is used to receive the collected data and perform real-time analysis of the status of the standby automatic transfer device. At the same time, it performs abnormal prediction and early warning based on the power consumption parameters of the station and the power consumption parameters of the equipment to determine whether there are potential fault hazards. The remote control module, connected to the status analysis and prediction module, is used to receive remote commands and remotely enable or disable the automatic transfer switch based on the remote commands. The data transmission and management module is connected to the multi-source data acquisition module, the status analysis and prediction module, and the remote control module. It is used to transmit the acquired data and analysis results to the monitoring room using the existing DTU equipment in the station, and to record and save the data completely and in a standardized manner. It is also used to send early warning information to the monitoring room through the data transmission and management module when the status analysis and prediction module determines that there is an anomaly.

2. The integrated online status monitoring device for power distribution rooms according to claim 1, characterized in that, The multi-source data acquisition module includes: The standby automatic transfer status acquisition submodule is used to collect charging status and operating status information of the standby automatic transfer device. This submodule establishes a data connection with the standby automatic transfer device through its communication interface and reads the charging flag and switch position signal data stored internally by the device according to a preset communication protocol. The charging status is represented by the character "C," which is set to "1" when charging is complete and "0" when not charging. The operating status is represented by the character "O," which is set to "1" during normal operation and "0" during an abnormality. The collected data is then subjected to CRC verification; if the verification fails, the data is collected again. The station building power consumption acquisition submodule is used to collect power consumption parameters of the station building. This submodule utilizes high-precision voltage and current transformers to convert the high voltage and large current of the station building's AC power consumption into small signals. AC voltage is represented by the character "VAC" in volts, and AC current is represented by the character "IAC" in amperes. During the acquisition process, the signals are sampled multiple times and noise interference is removed using a digital filtering algorithm. Finally, a timestamp is appended to the collected data. The equipment power consumption acquisition submodule is used to collect equipment power consumption parameters. It uses a Hall effect sensor to measure DC current and a voltage divider circuit to acquire DC voltage. DC voltage is represented by the character "VDC" in volts, and DC current is represented by the character "IDC" in amperes. The sensor and acquisition circuit are electromagnetically shielded in hardware, and the collected data is then verified and corrected. After data acquisition, the data is normalized and mapped to the [0, 1] interval.

3. The integrated online status monitoring device for power distribution rooms according to claim 2, characterized in that, The state analysis and prediction module performs online real-time analysis of the received data based on time series data, setting fixed time intervals. Within each time interval, the following data was collected by the multi-source data acquisition module: The charging status information and operating status information of the automatic transfer switch are represented by the characters C and O, respectively. Station building electrical parameters, including AC voltage and alternating current ; Equipment electrical parameters, including DC voltage and DC current Integrate and process.

4. The integrated online status monitoring device for power distribution rooms according to claim 3, characterized in that, For an automatic transfer switch (ATS) device, if the charging state C changes from "0" to "1" within n consecutive time intervals, and the operating state O remains "1", then the ATS device is determined to be in a normal charging and ready state. If C remains "0" and O remains "0", then the ATS device may be faulty, and the fault time sequence is recorded. .

5. The integrated online status monitoring device for power distribution rooms according to claim 4, characterized in that, An anomaly prediction and early warning system is employed using a machine learning-based long short-term memory network model; the collected data are organized into feature vectors according to their chronological order. , where t represents the time step; The core formula of the machine learning-based Long Short-Term Memory (LSTM) network model is: Forgotten Gate This is used to determine which information from the previous moment needs to be retained in the current moment. For the Sigmoid function, It is the weight matrix of the forget gate. It is the hidden state from the previous moment. It is the bias vector of the forget gate; Input gate , It is used to determine how new information at the current moment is added to the cell state. , It is the corresponding weight matrix. , is the bias vector, and tanh is the hyperbolic tangent function; Cell state Update cell state; Output gate Calculate the hidden state and output at the current moment.

6. The integrated online status monitoring device for power distribution rooms according to claim 5, characterized in that, During the model training phase, a large amount of historical data is used for training, including normal operation data and various failure data. Cross-validation is employed, dividing the dataset into k non-overlapping subsets, and each time... The model is trained using one subset and validated using the remaining subset. This process is repeated k times, and the average validation result is used as the model performance evaluation metric.

7. The integrated online status monitoring device for power distribution rooms according to claim 6, characterized in that, During model training, the hyperparameters of the machine learning-based Long Short-Term Memory (LSTM) network model are continuously adjusted to minimize the loss function on the validation set. The loss function uses mean squared error; the formula is as follows. Where N is the number of samples, It is the actual value. It is the predicted value; when the confidence level of the prediction result is lower than the set threshold... In such cases, abnormal situations are confirmed through manual intervention and further data analysis.

8. The integrated online status monitoring device for power distribution rooms according to claim 7, characterized in that, The state analysis and prediction module is also used to analyze the correlation between data, introducing a data correlation degree calculation formula to judge the rationality of the data; let the data correlation degree R be the relationship between the charging state C of the standby automatic transfer device and the AC voltage. The formula for calculating the degree of correlation is: ， Where n is the number of samples collected. and These represent the charging status and AC voltage data of the standby automatic transfer device collected in the i-th iteration. and These are the average values ​​of the charging status and AC voltage data, respectively; under normal circumstances... It should be within a reasonable range Inside; If the calculation yields Beyond this range, and AC voltage If the number of times the error exceeds a certain threshold m, it is suspected that the AC voltage data may be unreasonable due to sensor error. When it is suspected that a certain data is unreasonable due to sensor error, it is compared with the data of other related sensors. If the other data are all within the normal range and the difference from the current abnormal data is large, then the currently collected data is determined to be unreasonable data. The alarm of abnormal status of the backup automatic transfer device caused by the unreasonable data is a false alarm, and no alarm signal is issued. The abnormal data is recorded.

9. The integrated online status monitoring device for power distribution rooms according to claim 8, characterized in that, Before transmitting the analysis results to the monitoring room, the data transmission and management module verifies the results. For the analysis results of the automatic transfer switch (ATS) status, it checks again whether the actual status of the ATS matches the analysis results. If the analysis results indicate that the ATS should be activated, but the actual switch position shows it is disconnected and not charging, the analysis result is marked as suspicious. The data transmission and management module establishes a data interaction link with the data transmission and management modules of other devices within the station. When it receives relevant data from other devices, it comprehensively judges it with its own collected and analyzed data. When the device analyzes an abnormal power consumption of a device and simultaneously receives similar abnormal data from the data transmission and management modules of adjacent devices, it increases the credibility of its analysis results regarding the abnormal power consumption. If the data displayed by adjacent devices is normal, the device will check the accuracy of its own collected data again and recalculate the analysis results.

10. The integrated online status monitoring device for power distribution rooms according to claim 9, characterized in that, The data transmission and management module is also used to filter the collected data. It adopts a moving average filtering algorithm, sets a time window T, and calculates the moving average of the station power consumption parameters and equipment power consumption parameters within the time window T.

Images