Power distribution room comprehensive safety monitoring alarm system and fault prediction method

Abstract

The invention discloses a power distribution room comprehensive safety monitoring alarm system and a fault prediction method.The alarm system comprises a data acquisition module, a programmable logic controller and a communication module, and the data acquisition module comprises a temperature and humidity sensor, a smoke sensor, a water sensor and a gas monitoring sensor; the programmable logic controller comprises a power supply module, an analog quantity input module, a calculation and storage module, an analog quantity output module, a contactor, an indicating lamp and a local area network, the communication module comprises a power distribution room monitoring center and a wireless communication device, and the data acquisition module, the programmable logic controller and the communication module are connected in sequence. The fault prediction method comprises the following steps of data acquisition and preprocessing, feature extraction, model training and fault prediction and preventive maintenance decision making. The power distribution room safety monitoring mode is comprehensive, the automation degree of alarm is improved, the inspection pressure of operation and maintenance personnel is shared, and emergency safety events can be prevented or handled.

Description

Technical Field

[0001] The present invention belongs to the technical field of electrical engineering and automation, and particularly relates to a comprehensive safety monitoring and alarm system for a distribution room and a fault prediction method. Background Art

[0002] With the continuous development of the power industry, as an important part of the power system, the safety of the distribution room has been increasingly emphasized. Traditional monitoring of the distribution room mainly relies on manual inspections, which have problems such as low efficiency and poor accuracy. In order to improve the safety and reliability of the distribution room, various automated monitoring and alarm systems have emerged in recent years.

[0003] These systems usually include sensors such as temperature, humidity, and smoke, which are used to monitor the environmental parameters in the distribution room in real time and trigger alarms in abnormal situations. However, existing monitoring and alarm systems often have a single function and lack comprehensive safety assessment capabilities. In existing systems, the arrangement and selection of sensors are often not reasonable enough, resulting in insufficient accuracy and timeliness of monitoring data. For example, some key areas may not be fully covered, or the sensors used have low sensitivity and cannot detect potential safety hazards in a timely manner.

[0004] In terms of data processing and analysis, existing systems lack the ability to deeply mine and intelligently analyze complex data, which may lead to false alarms or missed alarms.

[0005] In the alarm mechanism, existing systems usually adopt the form of local audible and visual alarms, wired local area monitoring alarms, wireless network background system alarms alone or in combination, which require a high level of human inspection and monitoring panel operation, and lack an automated monitoring and instant alarm system, especially an alarm method for instant communication mobile terminals such as mobile phones. This makes it impossible for users to achieve all-time, instant, and efficient safety monitoring and emergency response in the distribution room. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a comprehensive safety monitoring and alarm system for a distribution room and a fault prediction method that can significantly improve the early warning accuracy rate of the system, shorten the early warning response time, and improve the intelligent and automated level of distribution room operation and maintenance.

[0007] The technical solution of the present invention is as follows: A comprehensive safety monitoring and alarm system for a distribution room, including: a data acquisition module, a programmable logic controller, a communication module, and an alarm module. The data acquisition module includes: a temperature and humidity sensor, a smoke sensor, a water immersion sensor, a gas monitoring sensor, and a partial discharge detector. The programmable logic controller includes: a power supply module, an analog input module, a calculation and storage module, an analog output module, a contactor, an indicator light, and a local area network. The communication module includes: a distribution room monitoring center and a wireless communication device. The alarm module includes: a short message notification, a telephone notification, and an audible and visual alarm; The data acquisition module is connected to the analog input module. The power supply module is respectively connected to the analog input module, the calculation and storage module, and the analog output module. The analog input module is connected to the calculation and storage module. The calculation and storage module is connected to the analog output module. The analog output module is respectively connected to the contactor and the indicator light. The analog input module and the analog output module are connected to the wireless communication device through the local area network. The wireless communication device is respectively wirelessly connected to the distribution room monitoring center and the alarm module. The distribution room monitoring center includes: a distribution room host, a platform server, and a data server.

[0008] Further, the temperature and humidity sensor is a digital sensor and is arranged in the distribution room.

[0009] Further, the smoke sensor is a photoelectric sensor and is arranged on the top of the distribution room.

[0010] Further, the water immersion sensor is arranged on the ground or low-lying area of the distribution room.

[0011] Further, the wireless communication device uses GPRS, 3G, 4G, or 5G communication for data transmission.

[0012] Further, the distribution room monitoring center is wirelessly connected to a mobile device through the wireless communication device.

[0013] Further, the analog output module is connected to the fire protection system of the distribution room.

[0014] A fault prediction method for a comprehensive safety monitoring and alarm system of a distribution room includes the following steps:

[0015] (1), Data acquisition and preprocessing, including the following steps:

[0016] S1, Data acquisition: Collect the following data through the temperature and humidity sensor, smoke sensor, water immersion sensor, and gas monitoring sensor in the system: temperature data, humidity data, smoke concentration data, water immersion situation data, and gas detection data;

[0017] S2. Data cleaning and processing: The data of each sensor generates time series data according to the sampling time. First, normalization processing is performed on the data of each sensor; for the normalized data, Kalman filtering is used to remove noise and fuse the information of multiple sensors; then, further smoothing and elimination are performed through moving average and three - times standard deviation to remove outliers; the time series data is aligned to ensure that data in each dimension is analyzed subsequently under the same timestamp.

[0018] (2). Feature extraction: Key features are extracted from the time series data. The key features include: mean, standard deviation, autocorrelation coefficient, frequency - domain features, and environmental compensation factors.

[0019] (3). Model training and fault prediction, including the following steps:

[0020] S3. Select the LSTM deep - learning model to process the time series data.

[0021] S4. Divide the sensor data into a training set and a test set, and optimize the model parameters through the mean square error.

[0022] S5. Use the trained model to predict the future trend of sensor data. When the difference between the predicted output value and the actual value exceeds the adaptive threshold, the system will issue a warning to indicate the occurring fault.

[0023] S6. Online learning optimization: Periodically detect whether there are false - alarm samples and perform incremental training to dynamically update the model parameters.

[0024] (4). Multi - modal alarm decision - making, including:

[0025] S7. On the basis of obtaining the predicted anomaly score or risk score in the fault prediction link, introduce a risk matrix to comprehensively evaluate multiple monitoring variables such as temperature and SF6 gas.

[0026] S8. Use fuzzy logic and a hierarchical alarm strategy, and can automatically execute corresponding emergency measures in combination with the actual situation.

[0027] (5). Preventive maintenance and self - learning, including:

[0028] S9. When the alarm level is relatively low, give maintenance suggestions, including: cleaning the ventilation equipment, checking the drainage system, and moisture - proof measures.

[0029] S10. When the alarm level is relatively high, trigger rapid emergency measures, including immediately cutting off the power supply and recording the log.

[0030] S11. All monitoring, alarm information, and false - alarm correction information are stored in the database for subsequent incremental training and model optimization.

[0031] S12. Perform self-learning regularly to further reduce the false alarm rate and improve the accuracy of fault prediction;

[0032] (6). Real-time communication and fault recovery, including:

[0033] S13. Adopt a lightweight custom communication protocol, including CRC checksum and reconnection in case of disconnection;

[0034] S14. Upload the important data after sensor fusion to the server after compression and verification, and the server will push alarm or maintenance messages to the corresponding terminals after parsing;

[0035] S15. When a network fault or server fault occurs, the system will automatically reconnect or switch to the local backup server.

[0036] Furthermore, the normalization process in step S2 is min-max normalization, and the expression is:

[0037]

[0038] where min(x i ), max(x i ) respectively represent the minimum and maximum values in the current window of the i-th sensor,

[0039] represents the sensor data, represents the normalized data;

[0040] The expressions for using Kalman filter to remove noise and fuse information from multiple sensors in step S2 include:

[0041] S21. Prediction stage:

[0042]

[0043] P k|k-1 = AP k-1|k-1 A T + Q,

[0044] S22. Update stage:

[0045] K k = P k|k-1 H T (HP k|k-1 H T + R) -1

[0046]

[0047] P k|k = (I - K k H)P k|k-1,

[0048] Among them, represent the prediction state and the update state respectively, and P k|k-1 , P k|k represent the covariance during prediction and update respectively, K k represents the Kalman gain, A, B, and H all represent measurement matrices, which are modeled based on each sensor, Q represents the process noise, R represents the measurement noise, z k represents the measured value, T represents the matrix transpose, and I represents the identity matrix;

[0049] The expression of the moving average in step S2 is:

[0050]

[0051] Among them, represents the data after smoothing processing, N represents the size of the moving window, and X(t - i) represents the original value of the time series data at time t - i;

[0052] The expression of the three - standard - deviation method is:

[0053] |X(t) - μ| > 3σ,

[0054] Among them, X(t) represents the value of the time series data at time t, μ represents the mean value of the time series data, and σ represents the standard deviation of the time series data;

[0055] The expression of the mean value is:

[0056]

[0057] Among them, X(t) represents the value of the time series data at time t, and T represents the number of samples;

[0058] The expression of the standard deviation is:

[0059]

[0060] Among them, X(t) represents the value of the time series data at time t, T represents the number of samples, and μ represents the mean value of the time series data;

[0061] The expression of the autocorrelation coefficient is:

[0062]

[0063] Among them, ACF( k) represents the autocorrelation coefficient at time lag k, whose range is [-1, 1]. X(t) represents the value of the time series data at time t, T represents the number of samples, μ represents the mean of the time series data, and k represents the time lag step;

[0064] The expression for the frequency domain feature is:

[0065]

[0066] where X(t) represents the value of the time series data at time t, and T represents the number of samples;

[0067] The expression for the LSTM deep learning model in step S3 is:

[0068] Input gate:

[0069] i t = σ(W i · [h t-1 , x t + b i ),

[0070] where i t represents the output of the input gate, σ represents the Sigmoid activation function, W i represents the weight matrix of the input gate, h t-1 represents the hidden state at the previous time t - 1, x t represents the input data at the current time t, and b i represents the bias term of the input gate;

[0071] Forget gate:

[0072] f t = σ(W f · [h t-1 , x t + b f ),

[0073] where f t represents the output of the forget gate, σ represents the Sigmoid activation function, W f represents the weight matrix of the forget gate, h t-1 represents the hidden state at the previous time t - 1, x t represents the input data at the current time t, and b f represents the bias term of the forget gate;

[0074] Memory cell update:

[0075]

[0076] where C tRepresents the cell state of the LSTM model at the current time t, f t Represents the output of the forget gate, i t Represents the output of the input gate, C t-1 Represents the cell state of the LSTM model at the previous time t-1, Represents the candidate cell state of the LSTM model at the current time t, h t-1 Represents the hidden state at the previous time t-1, x t Represents the input data at the current time t, W C Represents the weight matrix of the candidate cell state, b C Represents the bias term of the candidate cell state;

[0077] Output gate:

[0078] o t =σ(W o ·[h t-1 , x t +b o ),

[0079] where, o t Represents the output of the output gate, σ represents the Sigmoid activation function, W o Represents the weight matrix of the output gate, h t-1 Represents the hidden state at the previous time t-1, x t Represents the input data at the current time t, b o Represents the bias term of the output gate;

[0080] The expression of the mean square error in step S4 is:

[0081]

[0082] where, y i Represents the true value, Represents the predicted value, n represents the number of samples.

[0083] Furthermore, the adaptive threshold in step S5 is provided by the dynamic threshold adaptive algorithm, and the dynamic threshold adaptive algorithm includes:

[0084] S51. Basic threshold update:

[0085] θ base ←α·F t +(1-α)·θ base ,

[0086] where, θ base Represents the basic threshold, α represents the smoothing coefficient, taking 0.1-0.3, F t Represents the eigenvalue at the current time t;

[0087] S52. Environmental compensation:

[0088] θ final = θ base × γ,

[0089] where θ final represents the final threshold, and γ represents the environmental compensation factor calculated based on temperature, humidity, smoke concentration, and water immersion measurement values;

[0090] The calculation formula for the environmental compensation factor γ is:

[0091]

[0092] where β represents the scaling coefficient, which is used to control the comprehensive influence intensity of all environmental variables on γ. The larger the β value, the more obvious the overall influence of the environmental variables; the smaller the value, the smaller the adjustment range of the threshold; ∑ i∈{T，H，S，W，...} represents the weighted sum of a set of environmental variables, including temperature T, humidity H, smoke concentration S, and water immersion measurement value W, to comprehensively evaluate the influence of the environment on the system; w i represents the weight of the i-th environmental variable, reflecting the importance of this variable to the environmental risk or compensation factor; φ i (·) represents the fuzzy membership function or risk scoring function of the i-th variable, x′ i represents the normalized value of the i-th sensor data, γ ∈ [0.8, 1.5]. When the environment is relatively harsh, i.e., γ > 1, it corresponds to increasing vigilance. When the environment is relatively safe, i.e., γ < 1, it corresponds to decreasing vigilance.

[0093] Advantages of the present invention:

[0094] 1. Improve the accuracy and real-time performance of monitoring: By optimizing the sensor layout and selection, the present invention significantly improves the accuracy and real-time performance of monitoring data. Specifically, the improvement of sensor sensitivity enables the system to detect potential safety hazards more quickly, thereby reducing the risks of false alarms and missed alarms. Compared with the prior art, the present invention effectively reduces the false alarm rate;

[0095] 2. Enhance the data processing and analysis capabilities: The introduced advanced data processing and analysis algorithms enable the system to have a stronger intelligent early warning function. By constructing complex mathematical models and algorithms, the system can deeply mine and intelligently analyze the collected data, effectively improving the accuracy and timeliness of early warning, further increasing the early warning accuracy rate, and shortening the early warning response time;

[0096] 3. Implement a diversified alarm mechanism: The designed diversified alarm mechanism ensures that users can receive alarm information in a timely manner and take effective measures. Whether it is audible and visual alarms, SMS notifications, or phone voice messages, etc., they can be customized and adjusted according to user needs. This flexibility not only improves user satisfaction but also enhances the practicality and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not limit the embodiments of the present invention.

[0098] Figure 1 It is a schematic structural diagram of a comprehensive safety monitoring and alarm system for a power distribution room according to the present invention.

[0099] Figure 2 It is a schematic diagram of the usage state of a comprehensive safety monitoring and alarm system for a power distribution room according to the present invention.

[0100] Figure 3 It is a three-dimensional schematic diagram of a comprehensive safety monitoring and alarm system for a power distribution room according to the present invention.

[0101] In the figure, 1 - temperature and humidity sensor, 2 - smoke sensor, 3 - water immersion sensor, 4 - power distribution room host. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0102] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the embodiments and the drawings. The illustrative embodiments and descriptions thereof of the present invention are only used to explain the present invention and do not limit the present invention.

[0103] As Figure 1-2As shown in the figure, a comprehensive safety monitoring and alarm system for a distribution room includes a data acquisition module, a programmable logic controller, a communication module, and an alarm module. The data acquisition module includes a temperature and humidity sensor 1, a smoke sensor 2, a water immersion sensor 3, a gas monitoring sensor 4, and a partial discharge detector 5. The programmable logic controller includes a power supply module, an analog input module, a calculation and storage module (providing a real-time data visualization interface to display the current status of various monitoring indicators, storing historical data for convenient later query and analysis, identifying the trend of equipment status changes through data analysis tools, and providing a basis for subsequent maintenance decisions), an analog output module, a contactor, an indicator light, and a local area network. The communication module includes a distribution room monitoring center and a wireless communication device. The alarm module includes a short message notification (when the system detects an abnormal situation, it automatically sends a short message notification to a preset mobile phone number. The short message content includes a brief description of the abnormal situation, the occurrence time, and recommended countermeasures), a phone notification (when the system detects a serious abnormality or an emergency, it automatically dials a preset phone number for voice notification. The voice notification includes a detailed description of the abnormal situation, the occurrence time, and recommendations for emergency countermeasures), and an audible and visual alarm (when the system detects an abnormal situation, it immediately triggers the on-site audible and visual alarm device to remind on-site personnel through sound and light signals, so that the problem can be quickly located and preliminary countermeasures can be taken). The data acquisition module is connected to the analog input module. The power supply module is respectively connected to the analog input module, the calculation and storage module, and the analog output module. The analog input module is connected to the calculation and storage module. The calculation and storage module is connected to the analog output module. The analog output module is respectively connected to the contactor and the indicator light. The analog input module and the analog output module are connected to the wireless communication device through the local area network. The wireless communication device is wirelessly connected to the distribution room monitoring center and the alarm module respectively. The distribution room monitoring center includes a distribution room host 4, a platform server, and a data server.

[0104] Preferably, the temperature and humidity sensor 1 is a digital sensor and is installed in the distribution room.

[0105] Preferably, the smoke sensor 2 is a photoelectric sensor and is installed on the top of the distribution room.

[0106] Preferably, the water immersion sensor 3 is installed on the ground or in a low-lying area of the distribution room.

[0107] Preferably, the wireless communication device uses GPRS, 3G, 4G, or 5G communication for data transmission.

[0108] Preferably, the distribution room monitoring center is wirelessly connected to a mobile device (providing a mobile phone APP to support users to monitor the status of the distribution room anytime and anywhere) through the wireless communication device.

[0109] Preferably, the analog output module is connected to the fire protection system in the power distribution room.

[0110] A fault prediction method for an integrated safety monitoring and alarm system in a power distribution room includes the following steps:

[0111] (1) Data collection and preprocessing, including the following steps:

[0112] S1. Data collection: Collect the following data through the temperature and humidity sensors, smoke sensors, water immersion sensors, and gas monitoring sensors in the system: temperature data, humidity data, smoke concentration data, water immersion condition data, and gas detection data;

[0113] S2. Data cleaning and processing: The data of each sensor generates time series data according to the sampling time. First, perform normalization processing on the data of each sensor; for the normalized data, use the Kalman filter to remove noise and fuse the information of multiple sensors; then further smooth and eliminate through moving average and three times the standard deviation to remove outliers; align the time series data to ensure that the data of each dimension is analyzed under the same timestamp for subsequent analysis;

[0114] (2) Feature extraction: Extract key features from the time series data, and the key features include: mean value, standard deviation, autocorrelation coefficient, frequency domain features, and environmental compensation factors;

[0115] (3) Model training and fault prediction, including the following steps:

[0116] S3. Select the LSTM deep learning model to process the time series data;

[0117] S4. Divide the sensor data into a training set and a test set, and optimize the model parameters through the mean square error;

[0118] S5. Use the trained model to predict the future trend of sensor data. When the difference between the predicted output value and the actual value exceeds the adaptive threshold, the system will issue a warning to indicate the occurrence of a fault;

[0119] S6. Online learning optimization: Periodically detect whether there are false alarm samples and perform incremental training to dynamically update the model parameters;

[0120] (4) Multimodal alarm decision-making, including:

[0121] S7. On the basis of obtaining the predicted anomaly score or risk score in the fault prediction link, introduce a risk matrix to comprehensively evaluate multiple monitoring variables such as temperature and SF6 gas;

[0122] S8. Utilize fuzzy logic and a hierarchical alarm strategy (early warning, serious, emergency), and can automatically execute corresponding emergency measures (such as "immediately cut off power" or "only record logs", etc.) in combination with the actual situation;

[0123] (5). Preventive maintenance and self-learning, including:

[0124] S9. When the alarm level is relatively low, give maintenance suggestions, including: cleaning the ventilation equipment, checking the drainage system and moisture-proof measures;

[0125] S10. When the alarm level is relatively high, trigger rapid emergency measures, including immediately cutting off power and recording logs;

[0126] S11. All monitoring, alarm information, and false alarm correction information are stored in the database for subsequent incremental training and model optimization;

[0127] S12. Conduct self-learning regularly (such as weekly or monthly) to further reduce the false alarm rate and improve the accuracy of fault prediction;

[0128] (6). Real-time communication and fault recovery, including:

[0129] S13. Adopt a lightweight custom communication protocol, including CRC checksum and reconnection in case of disconnection;

[0130] S14. Compress and verify the important data after sensor fusion and upload it to the server, and the server will push alarm or maintenance messages to the corresponding terminals after parsing;

[0131] S15. When a network fault or server fault occurs, the system will automatically reconnect or switch to the local backup server.

[0132] Preferably, the normalization process in step S2 is min-max normalization, and the expression is:

[0133]

[0134] where min(x i ), max(x i ) respectively represent the minimum and maximum values in the current window of the i-th sensor, represents the sensor data, represents the normalized data;

[0135] The expression for using Kalman filter to remove noise and fuse information from multiple sensors in step S2 includes:

[0136] S21. Prediction stage:

[0137]

[0138] P k|k-1 = AP k-1|k-1 A T + Q,

[0139] S22. Update stage:

[0140] K k = P k|k-1 H T (HP k|k-1 H T + R) -1

[0141]

[0142] P k|k = (I - K k H)P k|k-1 ,

[0143] where, respectively represent the prediction state and the update state, P k|k-1 , P k|k respectively represent the covariance during prediction and update, K k represents the Kalman gain, A, B, and H all represent measurement matrices, modeled based on each sensor, Q represents process noise, R represents measurement noise, z k represents the measured value, T represents matrix transpose, and I represents the identity matrix;

[0144] The expression for the moving average in step S2 is:

[0145]

[0146] where, represents the data after smoothing, N represents the size of the moving window, and X(t - i) represents the original value of the time series data at time t - i;

[0147] The expression for the three - sigma method is:

[0148] |X(t)-μ|>3σ,

[0149] where, X(t) represents the value of the time series data at time t, μ represents the mean of the time series data, and σ represents the standard deviation of the time series data;

[0150] The expression for the mean is:

[0151]

[0152] where, X(t) represents the value of the time series data at time t, and T represents the number of samples;

[0153] The expression for the standard deviation is as follows:

[0154]

[0155] Where X(t) represents the value of the time series data at time t, T represents the number of samples, and μ represents the mean of the time series data;

[0156] The expression for the autocorrelation coefficient is as follows:

[0157]

[0158] Where ACF(k) represents the autocorrelation coefficient at time lag k, and its range is [-1, 1]. X(t) represents the value of the time series data at time t, T represents the number of samples, μ represents the mean of the time series data, and k represents the time lag step;

[0159] The expression for the frequency domain feature is as follows:

[0160]

[0161] Where X(t) represents the value of the time series data at time t, and T represents the number of samples;

[0162] The expression for the LSTM deep learning model in step S3 is as follows:

[0163] Input gate:

[0164] i t = σ(W i · [h t-1 , x t + b i ),

[0165] Where i t represents the output of the input gate, σ represents the Sigmoid activation function, W i represents the weight matrix of the input gate, h t-1 represents the hidden state at the previous time t - 1, x t represents the input data at the current time t, and b i represents the bias term of the input gate;

[0166] Forget gate:

[0167] f t = σ(W f · [h t-1 , x t + b f ),

[0168] Where f tdenotes the output of the forget gate, σ denotes the Sigmoid activation function, W f denotes the weight matrix of the forget gate, h t-1 denotes the hidden state at the previous time step t - 1, x t denotes the input data at the current time step t, b f denotes the bias term of the forget gate;

[0169] Memory cell update:

[0170]

[0171] where C t denotes the cell state of the LSTM model at the current time step t, f t denotes the output of the forget gate, i t denotes the output of the input gate, C t-1 denotes the cell state of the LSTM model at the previous time step t - 1, denotes the candidate cell state of the LSTM model at the current time step t, h t-1 denotes the hidden state at the previous time step t - 1, x t denotes the input data at the current time step t, W C denotes the weight matrix of the candidate cell state, b C denotes the bias term of the candidate cell state;

[0172] Output gate:

[0173] o t = σ(W o · [h t-1 , x t + b o ),

[0174] where o t denotes the output of the output gate, σ denotes the Sigmoid activation function, W o denotes the weight matrix of the output gate, h t-1 denotes the hidden state at the previous time step t - 1, x t denotes the input data at the current time step t, b o denotes the bias term of the output gate;

[0175] The expression for the mean squared error in step S4 is:

[0176]

[0177] where y i denotes the true value, denotes the predicted value, and n denotes the number of samples.

[0178] Preferably, the adaptive threshold in step S5 is provided by a dynamic threshold adaptive algorithm, and the dynamic threshold adaptive algorithm includes:

[0179] S51. Basic threshold update:

[0180] θ base ←α·F t +(1-α)·θ base ,

[0181] where θ base represents the basic threshold, α represents the smoothing coefficient, taking 0.1 - 0.3, and F t represents the eigenvalue at the current moment t;

[0182] S52. Environmental compensation:

[0183] θ final =θ base ×γ,

[0184] where θ final represents the final threshold, and γ represents the environmental compensation factor calculated based on temperature, humidity, smoke concentration, and water immersion measurement values;

[0185] The calculation formula of the environmental compensation factor γ is:

[0186]

[0187] where β represents the scaling coefficient, used to control the comprehensive influence intensity of all environmental variables on γ. The larger the β value, the more obvious the overall influence of the environmental variables; the smaller the value, the smaller the adjustment range of the threshold; ∑ i∈{T，H，S，W，...} represents the weighted sum of a set of environmental variables, including temperature T, humidity H, smoke concentration S, and water immersion measurement value W, to comprehensively evaluate the influence of the environment on the system; w i represents the weight of the i-th environmental variable, reflecting the importance of this variable to the environmental risk or compensation factor; φ i (·) represents the fuzzy membership function or risk scoring function of the i-th variable, and x′ i represents the normalized value of the i-th sensor data, γ ∈ [0.8, 1.5]. When the environment is relatively harsh, that is, γ > 1, it corresponds to raising vigilance. When the environment is relatively safe, that is, γ < 1, it corresponds to lowering vigilance.

[0188] The above specific implementation manners further elaborate on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above is only the specific implementation manners of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A comprehensive safety monitoring and alarm system for a power distribution room, characterized in that, Including: A data acquisition module, a programmable logic controller, a communication module, and an alarm module. The data acquisition module includes: a temperature and humidity sensor (1), a smoke sensor (2), a water immersion sensor (3), a gas monitoring sensor (4), and a partial discharge detector (5). The programmable logic controller includes: a power supply module, an analog input module, a calculation and storage module, an analog output module, a contactor, an indicator light, and a local area network. The communication module includes: a power distribution room monitoring center and a wireless communication device. The alarm module includes: a short message notification, a telephone notification, and an audible and visual alarm; The data acquisition module is connected to the analog input module. The power supply module is respectively connected to the analog input module, the calculation and storage module, and the analog output module. The analog input module is connected to the calculation and storage module. The calculation and storage module is connected to the analog output module. The analog output module is respectively connected to the contactor and the indicator light. The analog input module and the analog output module are connected to the wireless communication device through the local area network. The wireless communication device is wirelessly connected to the power distribution room monitoring center and the alarm module respectively. The power distribution room monitoring center includes: a power distribution room host (4), a platform server, and a data server.

2. The integrated safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that, The temperature and humidity sensor (1) is a digital sensor and is arranged in the power distribution room.

3. The integrated safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that The smoke sensor (2) is a photoelectric sensor and is arranged at the top of the power distribution room.

4. A comprehensive safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that, The water immersion sensor (3) is arranged on the ground or low-lying area of the power distribution room.

5. The integrated safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that, The wireless communication device uses GPRS, 3G, 4G, or 5G communication for data transmission.

6. The integrated safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that, The power distribution room monitoring center is wirelessly connected to a mobile device through the wireless communication device.

7. The integrated safety monitoring and alarm system for a power distribution room according to claim 1, characterized in that The analog output module is connected to the fire protection system of the power distribution room.

8. A fault prediction method for an integrated safety monitoring and alarm system in a power distribution room, characterized in that, Including the following steps: (1) Data acquisition and preprocessing, including the following steps: S1. Data acquisition: Collect the following data through the temperature and humidity sensor, smoke sensor, water immersion sensor, and gas monitoring sensor in the system: temperature data, humidity data, smoke concentration data, water immersion condition data, and gas detection data; S2. Data cleaning and processing: The data of each sensor generates time series data according to the sampling time. For each sensor data, first perform normalization processing; for the normalized data, use the Kalman filter to remove noise and fuse the information of multiple sensors; then further smooth and eliminate through moving average and three times the standard deviation to remove outliers; align the time series data to ensure that each dimension of data is analyzed under the same time stamp for subsequent analysis; (2) Feature extraction: Extract key features from the time series data. The key features include: mean value, standard deviation, autocorrelation coefficient, frequency domain features, and environmental compensation factor; (3) Model training and fault prediction, including the following steps: S3. Select the LSTM deep learning model to process the time series data; S4. Divide the sensor data into a training set and a test set, and optimize the model parameters through the mean square error; S5. Use the trained model to predict the future trend of sensor data. When the difference between the predicted output value and the actual value exceeds the adaptive threshold, the system will issue a warning to indicate the occurrence of a fault. S6. Online learning optimization: Periodically detect whether there are false alarm samples and perform incremental training to dynamically update the model parameters. (4). Multimodal alarm decision-making, including: S7. On the basis of obtaining the predicted anomaly score or risk score in the fault prediction stage, introduce a risk matrix to comprehensively evaluate multiple monitoring variables such as temperature and SF6 gas. S8. Use fuzzy logic and hierarchical alarm strategies, and can automatically execute corresponding emergency measures in combination with the actual situation. (5). Preventive maintenance and self-learning, including: S9. When the alarm level is relatively low, give maintenance suggestions, including: cleaning the ventilation equipment, checking the drainage system and moisture-proof measures. S10. When the alarm level is relatively high, trigger rapid emergency measures, including immediately powering off and recording logs. S11. All monitoring, alarm information, and false alarm correction information are stored in the database for subsequent incremental training and model optimization. S12. Perform self-learning regularly to further reduce the false alarm rate and improve the accuracy of fault prediction. (6). Real-time communication and fault recovery, including: S13. Adopt a lightweight custom communication protocol, including CRC checksum and disconnection reconnection. S14. Compress and verify the important data after sensor fusion and upload it to the server. After parsing, the server will push alarm or maintenance messages to the corresponding terminals. S15. When a network fault or server fault occurs, the system will automatically reconnect or switch to the local backup server.

9. The fault prediction method of a comprehensive safety monitoring and alarm system for a power distribution room according to claim 8, characterized in that, The normalization process in step S2 is min-max normalization, and the expression is: where min(x i ) and max(x i ) respectively represent the minimum and maximum values in the current window of the i-th sensor, represents the sensor data, represents the data after normalization; The expression for using Kalman filtering to remove noise and fuse information from multiple sensors in step S2 includes: S21. Prediction stage: P k|k-1 = AP k-1|k-1 A T + Q, S22. Update stage: K k = P k|k-1 H T (HP k|k-1 H T + R) -1 P k|k = (I - K k H)P k|k-1 , Among them, respectively represent the prediction state and the update state, P k|k-1 , P k|k respectively represent the covariance during prediction and update, K k represents the Kalman gain, A, B, and H all represent measurement matrices, modeled according to each sensor, Q represents the process noise, R represents the measurement noise, z k represents the measured value, T represents the matrix transpose, and I represents the identity matrix; The expression for moving average in step S2 is: Among them, represents the data after smoothing processing, N represents the size of the sliding window, and X(t - i) represents the original value of the time series data at time t - i; The expression for the three-sigma method is: |X(t) - μ| > 3σ where X(t) represents the value of the time series data at time t, μ represents the mean of the time series data, and σ represents the standard deviation of the time series data. The expression for the mean is: where X(t) represents the value of the time series data at time t, and T represents the number of samples. The expression for the standard deviation is: where X(t) represents the value of the time series data at time t, T represents the number of samples, and μ represents the mean of the time series data. The expression for the autocorrelation coefficient is: where ACF(k) represents the autocorrelation coefficient at time lag k, and its range is [-1, 1], X(t) represents the value of the time series data at time t, T represents the number of samples, μ represents the mean of the time series data, and k represents the time lag step. The expression for the frequency domain feature is: where X(t) represents the value of the time series data at time t, and T represents the number of samples. The expression for the LSTM deep learning model in step S3 is: Input gate: i t = σ(W i · [h t-1 , x t + b i ) where, i t represents the output of the input gate, σ represents the Sigmoid activation function, W i represents the weight matrix of the input gate, h t-1 represents the hidden state at the previous time step t - 1, x t represents the input data at the current time step t, b i represents the bias term of the input gate; Forget gate: f t = σ(W f · [h t-1 , x t + b f ), Among them, f t represents the output of the forgetting gate, σ represents the Sigmoid activation function, W f represents the weight matrix of the forgetting gate, h t-1 represents the hidden state at the previous moment t-1, x t represents the input data at the current moment t, b f represents the bias term of the forgetting gate; Memory cell update: Among them, C t represents the cell state of the LSTM model at the current time t, f t represents the output of the forget gate, i t represents the output of the input gate, C t-1 represents the cell state of the LSTM model at the previous time t-1, represents the candidate cell state of the LSTM model at the current time t, h t-1 represents the hidden state at the previous time t-1, x t represents the input data at the current time t, W C represents the weight matrix of the candidate cell state, b C represents the bias term of the candidate cell state; Output gate: o t = σ(W o · [h t-1 , x t + b o ), Among them, o t represents the output of the output gate, σ represents the Sigmoid activation function, W o represents the weight matrix of the output gate, h t-1 represents the hidden state at the previous time step t - 1, x t represents the input data at the current time step t, b o represents the bias term of the output gate; The expression for the mean squared error in step S4 is: Among them, y i represents the true value, represents the predicted value, and n represents the number of samples.

10. The fault prediction method of a comprehensive safety monitoring and alarm system for a power distribution room according to claim 8, characterized in that, The adaptive threshold in step S5 is provided by a dynamic threshold adaptive algorithm, and the dynamic threshold adaptive algorithm includes: S51. Basic threshold update: θ base ←α·F t +(1 - α)·θ base , where θ base represents the base threshold, α represents the smoothing coefficient, taking 0.1 - 0.3, and F t represents the eigenvalue at the current moment t; S52. Environmental compensation: θ final = θ base × γ, where θ final represents the final threshold, and γ represents the environmental compensation factor calculated based on temperature, humidity, smoke concentration, and water immersion measurement values; The calculation formula of the environmental compensation factor γ is: Among them, β represents the scaling factor, which is used to control the comprehensive influence intensity of all environmental variables on γ. The larger the β value, the more obvious the overall influence of the environmental variables; the smaller the value, the smaller the adjustment range for the threshold; ∑ i∈{T,H,S,W,...} represents the weighted sum of a group of environmental variables, including temperature T, humidity H, smoke concentration S, and water immersion measurement value W, to comprehensively evaluate the influence of the environment on the system; w i represents the weight of the i-th environmental variable, reflecting the importance of this variable to the environmental risk or compensation factor; φ i (·) represents the fuzzy membership function or risk scoring function of the i-th variable, x′ i represents the normalized value of the i-th sensor data, γ ∈ [0.8, 1.5]. When the environment is relatively harsh, that is, γ > 1, corresponding vigilance is increased; when the environment is relatively safe, that is, γ < 1, corresponding vigilance is decreased.

Images