A high-rise building dense insulation type bus duct overheating fault early warning system and method

Abstract

The present application relates to an overheating fault early warning system and method for densely insulated bus ducts in high-rise buildings. The system includes a gas sensor installed in a wiring room, the gas sensor is used to collect gas concentration data at the installation position of the densely insulated bus duct, the gas sensor is connected to a WIFI gateway, the WIFI gateway is connected to a local area network, the local area network is connected to a server, the gas sensor transmits the collected gas concentration data to the server through the WIFI gateway and the local area network, the server processes the gas concentration data, extracts its features, and inputs them into a trained AdaBoost model for discrimination. If an overheating fault is identified, an alarm message is pushed to the operation and maintenance personnel. The present application has a high recognition accuracy and can be widely used for overheating fault monitoring of densely insulated bus ducts in high-rise buildings.

Description

Technical Field

[0001] The present application relates to the field of disaster prevention for power distribution networks, and in particular to a system and method for early warning overheating faults of densely insulated bus ducts in high-rise buildings. Background Art

[0002] Densely insulated bus ducts have emerged as a new type of distribution conductor. Compared with traditional cables, they fully demonstrate their superiority when transmitting large currents. At the same time, due to the use of new processes and technologies, the contact resistance and temperature rise at the connection points at both ends of the bus duct and the branch port are greatly reduced, and high-quality insulating materials are used in the bus duct, thereby improving the safety and reliability of the bus duct and making the entire system more complete and reliable.

[0003] In high-rise buildings, with the rapid increase in electricity consumption, especially the emergence of numerous high-rise buildings and large factories and workshops, traditional cables as transmission conductors can no longer meet the requirements in high-current transmission systems. The parallel use of multiple cables has brought a lot of inconvenience to on-site installation and construction. Densely insulated bus ducts are widely used in the power supply and distribution of newly built high-rise buildings because of their strong heat dissipation capacity, flexible wiring, and strong anti-dynamic thermal stability.

[0004] Due to various reasons such as excessive current, poor factory quality, improper installation process, etc., the dense bus duct sometimes causes heating, especially at the joints. In the dense insulation bus duct, the bus conductor copper bar is wrapped with a layer of PET insulation. When the overheating fault lasts for a period of time, the high temperature will cause physical damage to the insulation layer, resulting in a decrease in insulation performance, which can easily lead to short circuits, fires or explosions, power outages and economic losses.

[0005] As the temperature of the insulation layer of the dense insulation bus duct increases, it can be roughly divided into four stages: A, normal stage, B, overheating stage (no smoke), C, melting stage, D, burning or explosion stage. The purpose of the present invention is to detect faults in the overheating stage, thereby sending an alarm signal to the operation and maintenance personnel to prevent short circuit faults and fires.

[0006] In the current early identification systems for electrical equipment fires, most of the research targets are PVC cables, and the research focuses on the smoldering (melting) stage and the burning stage. These early warning systems identify fires by collecting data from smoke sensors, gas sensors, video images, infrared images, ambient temperature, and other data. Although this type of early fire warning system can detect fires in the smoldering stage or smoke stage of electrical fires, the early stage here refers to the early stage of fire generation, provided that the fire has already occurred. There are two points where the present invention differs from these studies (systems): First, the research target is the densely insulated bus duct used in high-rise buildings. Second, it detects the overheating fault stage (fire has not yet occurred), rather than the early stage of fire.

[0007] The structure of densely insulated bus duct is very compact, and it is not convenient to install thermocouples inside the bus duct to measure its temperature on site. If a thermocouple is installed on the bus conductor, its insulation structure will be destroyed. When the bus in the bus duct heats up, the temperature of the corresponding shell part is higher than other parts of the other shells, so it can be measured with an infrared thermometer. In order to find overheating faults of densely insulated bus ducts, the bus duct can be regularly inspected by thermal imaging equipment such as infrared imagers during inspections to find overheating points. However, the space in the wiring room of high-rise buildings is usually small, which is not conducive to infrared detection. Secondly, this kind of inspection wastes a lot of manpower and material resources, and overheating faults cannot be found in real time. If infrared imagers are arranged along the bus duct in the wiring room for real-time monitoring, the cost is very high. Summary of the invention

[0008] The purpose of the embodiments of the present application is to provide a high-rise building dense insulation type bus duct overheating fault early warning system and method, which has high recognition accuracy and can be widely used in high-rise building dense insulation type bus duct overheating fault monitoring.

[0009] To achieve the above objectives, this application provides the following technical solutions:

[0010] In the first aspect, an embodiment of the present application provides a high-rise building dense insulation bus duct overheating fault warning system, including a gas sensor installed in a wiring room, the gas sensor is used to collect gas concentration data at the installation position of the dense insulation bus duct, the gas sensor is connected to a WIFI gateway, the WIFI gateway is connected to a local area network, the local area network is connected to a server, the gas sensor transmits the collected gas concentration data to the server through the WIFI gateway and the local area network, the server processes the gas concentration data, extracts its features, and inputs them into a trained AdaBoost model for judgment. If an overheating fault is identified, an alarm message is pushed to the operation and maintenance personnel.

[0011] The gas sensor transmits the collected data to the WIFI gateway via WIFI wireless transmission.

[0012] The server pushes the alarm information to the operation and maintenance personnel via SMS or applet.

[0013] In a second aspect, an embodiment of the present application provides a method for early warning of overheating faults of densely insulated bus ducts in high-rise buildings, characterized in that the method uses the above-mentioned early warning system for overheating faults of densely insulated bus ducts in high-rise buildings, and the method includes the following specific steps:

[0014] An experimental environment was built, and the bus duct was placed in a closed glass cover to simulate a building wiring room. Then, a continuously increasing high current was applied to the bus, and the temperature of the bus copper bar also increased accordingly. The characteristic gas concentrations in the normal state and the overheated state were collected and marked, and then their features were extracted to form a sample data set. The data set was randomly divided into a test set and a training set. The test set and the training set were used to optimize the parameters of the AdaBoost classifier to obtain the trained AdaBoost overheating judgment model.

[0015] The gas sensors installed in the building wiring room collect VOCs gas concentration in real time, once every 2 seconds. After the collected data is transmitted back to the server, the received VOCs gas concentration time series data is first feature extracted and then input into the trained AdaBoost overheating judgment model. If the model outputs an overheating fault, the alarm information is sent to the operation and maintenance personnel.

[0016] The feature extraction of the VOCs gas concentration time series data is specifically as follows:

[0017] For the gas concentration data time series x i , and perform the maximum value Max, minimum value Min, average value Avg and variance feature δ 2 The specific formula is as follows:

[0018]

[0019]

[0020]

[0021]

[0022] Where n is the length of the vector, then a time series G with a length of n has a characteristic F G , which is expressed as follows:

[0023] F G ={Max(G), Min(G), Avg(G), Var(G)} (5).

[0024] Compared with the prior art, the beneficial effects of the present application are as follows: the present application designs an overheating fault identification alarm system for the overheating fault detection problem of the densely insulated bus duct widely used in high-rise buildings. The principle is to analyze the gas generated in the early stage of pyrolysis of the bus duct insulation layer, explore the relationship between its concentration and the temperature of the bus insulation layer, and find the characteristic gas. Then design a sensor to collect the gas concentration, and convert the overheating fault detection problem into a binary classification problem. Samples are obtained through experiments, and feature extraction is performed. The AdaBoost machine learning classification model is used for testing and training to optimize the model parameters. In the application, the corresponding characteristic gas sensors are arranged on-site in the wiring room of the building, and the collected sensor data is transmitted to the server side through the WIFI network. The trained AdaBoost model is used on the server side to input the sensor data to judge the overheating fault state of the bus duct. Here, the on-site state is defined as 2 categories: 1. Normal state; 2. Overheating fault state. The example shows that the recognition accuracy of the bus duct overheating fault alarm system is high. The alarm system designed in the present application can be widely used for overheating fault monitoring of densely insulated bus ducts in high-rise buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 Schematic diagram of the system structure of an embodiment of the present application.

[0027] Figure 2 Schematic diagram of the method flow of an embodiment of the present application.

[0028] Figure 3 This is a diagram of classification and recognition accuracy of the AdaBoost algorithm with different parameters in the embodiment of the present application.

[0029] Figure 4 This is a comparison chart of the accuracy of different classification algorithms in the embodiments of the present application. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0031] The terms "comprises," "comprising," or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0032] The busbar conductor of the densely insulated bus duct is a copper bar, which is wrapped with a layer of PET (polyester film) insulation material. Polyester film is made of polyethylene terephthalate as raw material, extruded into thick sheets, and then biaxially stretched to form a film material. It is a colorless, transparent, and glossy film with excellent performance, high rigidity, hardness and toughness, puncture resistance, friction resistance, high and low temperature resistance, chemical resistance, oil resistance, good air tightness and freshness preservation. It is one of the commonly used barrier composite film substrates.

[0033] When PET is used as a bus duct insulation material, in order to improve its insulation, stability and flame retardancy, many additives are added during the production process of the bus duct insulation material. Moreover, the ingredients of these additives are kept confidential by the manufacturer due to commercial needs. The ingredients of different manufacturers are slightly different.

[0034] When the insulation layer in the densely insulated bus duct is heated and the temperature reaches a certain level, different products will be decomposed at different temperatures. A gas with high concentration and easy to detect can be selected as its characteristic gas. Then, by detecting the concentration of the characteristic gas produced in the overheating stage, early warning of overheating faults of the bus duct can be achieved.

[0035] like Figure 1 As shown, a high-rise building dense insulation bus duct overheating fault early warning system includes a gas sensor 2 installed in a wiring room, the gas sensor 2 is used to collect gas concentration data at the installation position of the dense insulation bus duct 1, the gas sensor 2 is connected to a WIFI gateway 3, the WIFI gateway is connected to a local area network 4, the local area network 4 is connected to a server 5, the gas sensor 2 transmits the collected gas concentration data to the server 5 through the WIFI gateway and the local area network 4, the server 5 processes the gas concentration data, extracts its features, and inputs them into a trained AdaBoost model for judgment. If an overheating fault is identified, an alarm message is pushed to an operation and maintenance personnel 6.

[0036] The gas sensor 2 transmits the collected data to the WIFI gateway via WIFI wireless transmission.

[0037] The server 5 pushes the alarm information to the operation and maintenance personnel 6 via SMS or mini-program.

[0038] A method for early warning of overheating faults of densely insulated bus ducts in high-rise buildings, using the above-mentioned early warning system for overheating faults of densely insulated bus ducts in high-rise buildings, the method comprises the following specific steps:

[0039] An experimental environment was built, and the bus duct was placed in a closed glass cover to simulate a building wiring room. Then, a continuously increasing high current was applied to the bus, and the temperature of the bus copper bar also increased accordingly. The characteristic gas concentrations in the normal state and the overheated state were collected and marked, and then their features were extracted to form a sample data set. The data set was randomly divided into a test set and a training set. The test set and the training set were used to optimize the parameters of the AdaBoost classifier to obtain the trained AdaBoost overheating judgment model.

[0040] The gas sensor 2 installed in the building wiring room collects VOCs gas concentration in real time, once every 2 seconds. After the collected data is transmitted back to the server, the received VOCs gas concentration time series data is firstly extracted for features, and then input into the trained AdaBoost overheating judgment model. If the model outputs an overheating fault, the alarm information is sent to the operation and maintenance personnel.

[0041] The feature extraction of the VOCs gas concentration time series data is specifically as follows:

[0042] For the gas concentration data time series x i , and perform the maximum value Max, minimum value Min, average value Avg and variance feature δ 2 The specific formula is as follows:

[0043]

[0044]

[0045]

[0046]

[0047] Where n is the length of the vector, then a time series G with a length of n has a characteristic F G , which is expressed as follows:

[0048] F G ={Max(G), Min(G), Avg(G), Var(G)} (5).

[0049] The application process on the server side is as follows:

[0050] 1) Receive a data and put it into an array with a fixed length of 5;

[0051] 2) If the array is full of 5 items, proceed to the next step;

[0052] 3) Extract the features of the array according to formula 1-5 and clear the array to zero;

[0053] 4) Input the extracted features into the trained AdaBoost judgment model;

[0054] 5) If the model output is 1 (overheating state), an alarm message is issued; if the model output is 0 (normal state), continue with step 1).

[0055] AdaBoost classification model,

[0056] AdaBoost is the abbreviation of "Adaptive Boosting". Its adaptability lies in that the weights of samples that were misclassified by the previous basic classifier will increase, while the weights of samples that were correctly classified will decrease and be used again to train the next one. In other words, weak classifiers with low error rates have a larger weight in the final classifier, otherwise they will have a smaller weight than the basic classifier. At the same time, in each round of iteration, a new weak classifier is added until a predetermined sufficiently small error rate is reached or the pre-specified maximum number of iterations is reached to determine the final strong classifier.

[0057] The algorithm flow is as follows:

[0058] Input: training data set T, T = {(x1, y1), (x2, y2), ..., (x N ,y N )},in, y i ∈Y={-1,+1}, the number of iterations is M.

[0059] Initialize the weight distribution of the training samples, that is

[0060]

[0061] For m = 1, 2, ... M

[0062] (a) Using a weighted distribution D m The training data set is used to learn and obtain the weak classifier G m (x)

[0063] (b) Calculate G m (x) Classification error rate on the training data set:

[0064]

[0065] (c) Calculate G m (x) The weight of the strong classifier:

[0066]

[0067] (d) Update the weight distribution of the training dataset (here z m is the normalization factor, in order to make the probability distribution of the sample equal to 1)

[0068]

[0069]

[0070] Get the final classifier

[0071]

[0072] Model training and testing

[0073] 80% of the collected 3000 sample feature sets are used as training sets and 20% as test sets. After training and testing, the AdaBoost model parameters are solved and optimized.

[0074] Here AdaBoost is implemented using the open source Scikit-Learn library, and the classification function is AdaBoostClassifier(), whose main input parameters are n_estimators (positive integer) and learning_rate (floating point number). Among them, n_estimators is the number of base classifier boosts (cycles), and learning_rate is the learning rate, which indicates the gradient convergence speed.

[0075] Different n_estimators and different learning_rate(C), the recognition accuracy is as follows Figure 3 As shown in Figure 1, when n_estimators = 200 and C = 0.2, the highest average accuracy is 89.2%. Its distribution is shown in Table 1.

[0076] Table 1. Classification accuracy

[0077]

[0078] Then the parameters in this model are determined as: n_estimators = 200, learning_rate = 0.2.

[0079] The AdaBoost classification algorithm used in the present invention is compared with other commonly used machine learning classification algorithms, such as K-nearest neighbor (KNN), support vector machine (SVM), naive Bayes, decision tree, random forest, etc., and the recognition accuracy is compared. Figure 4 ,from Figure 4 It can be seen that in the present invention, the fault classification is performed based on the concentration of characteristic gas of bus duct overheating, and the AdaBoost classification algorithm is used, and its classification and recognition accuracy is higher than that of other classification algorithms.

[0080] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for early warning of overheating fault of densely insulated bus duct in high-rise buildings, characterized in that: A high-rise building dense insulation type bus duct overheating fault early warning system is used. The system includes a gas sensor (2) installed in a wiring room. The gas sensor (2) is used to collect gas concentration data at the installation position of the dense insulation type bus duct (1). The gas sensor (2) is connected to a WIFI gateway (3), the WIFI gateway is connected to a local area network (4), and the local area network (4) is connected to a server (5). The gas sensor (2) transmits the collected gas concentration data to the server (5) through the WIFI gateway and the local area network (4). The server (5) processes the gas concentration data, extracts its features, and inputs them into a trained AdaBoost model for discrimination. If an overheating fault is identified, an alarm message is pushed to an operation and maintenance personnel (6). The method includes the following specific steps: An experimental environment was built, and the bus duct was placed in a closed glass cover to simulate a building wiring room. Then, a continuously increasing high current was applied to the bus, and the temperature of the bus copper bar also increased accordingly. The characteristic gas concentrations in the normal state and the overheated state were collected and marked, and then their features were extracted to form a sample data set. The data set was randomly divided into a test set and a training set. The test set and the training set were used to optimize the parameters of the AdaBoost classifier to obtain the trained AdaBoost overheating judgment model. The gas sensor (2) installed in the building wiring room collects VOCs gas concentration in real time, once every 2 seconds. After the collected data is transmitted back to the server, the received VOCs gas concentration time series data is firstly subjected to feature extraction, and then input into the trained AdaBoost overheating judgment model. If the model output indicates that an overheating fault is found, an alarm message is sent to the operation and maintenance personnel.

2. A high-rise building dense insulation type bus duct overheating fault early warning method according to claim 1, characterized in that: The feature extraction of the VOCs gas concentration time series data is specifically as follows: For the gas concentration data time series x i , and perform the maximum value Max, minimum value Min, average value Avg and variance feature δ 2 The specific formula is as follows: Where n is the length of the vector, then a time series G with a length of n has a characteristic F G , which is expressed as follows: F G ={Max(G),Min(G),Avg(G),Var(G)} (5).

3. The method for early warning of overheating fault of densely insulated bus duct in high-rise buildings according to claim 1 is characterized in that: The gas sensor (2) transmits the collected data to the WIFI gateway by means of WIFI wireless transmission.

4. The method for early warning of overheating fault of densely insulated bus duct in high-rise buildings according to claim 1 is characterized in that: The server (5) pushes the alarm information to the operation and maintenance personnel (6) via SMS or mini-program.

Images