Fault Detection and Identification of Furnace Negative Pressure System with CVA and GA-XGBoost

Abstract

The boiler is an essential energy conversion facility in a thermal power plant. One small malfunction or abnormal event will bring huge economic loss and casualties. Accurate and timely detection of abnormal events in boilers is crucial for the safe and economical operation of complex thermal power plants. Data-driven fault diagnosis methods based on statistical process monitoring technology have prevailed in thermal power plants, whereas the false alarm rates of those methods are relatively high. To work around this, this paper proposes a novel fault detection and identification method for furnace negative pressure system based on canonical variable analysis (CVA) and eXtreme Gradient Boosting improved by genetic algorithms (GA-XGBoost). First, CVA is used to reduce the data redundancy and construct the canonical residuals to measure the prediction ability of the state variables. Then, the fault detection model based on GA-XGBoost is schemed using the constructed canonical residual variables. Specially, GA is introduced to determine the optimal hyperparameters of XGBoost and speed up the convergence. Next, this paper presents a novel fault identification method based on the reconstructed contribution statistics, considering the contribution of state space, residual space and canonical residual space. Besides, the proposed statistics renders different weights to the state vectors, the residual vectors and the canonical residual vectors to improve the sensitivity of faulty variables. Finally, the real industrial data from a boiler furnace negative pressure system of a certain thermal power plant is used to demonstrate the ability of the proposed method. The result demonstrates that this method is accurate and efficient to detect and identify the faults of a true boiler.

Research on the Damage Diagnosis Model Algorithm of Cable-Stayed Bridges Based on Data Mining

“…The values of the gamma and c parameters of the SVR model are the same as those of the SVC model, epsilon=np.linspace (0.00001,1,30); the XGBClassifier model parameter value range is n_estimators=np.arange(50,401,30), max_depth=np.arange(2,10,1), min_chile_Weight=np.linspace (1,9,30) and learning_rate= np.linspace(0.05,0.3). The XGBRegenerator model parameter value range is n_estimators= np.arange(50,401,10), max_depth=np.arange (1,15), min_chile_weigh=np.linspace(1,9,50), learning_rate=np.linspace(0.05,0.3,50). Among these, np.linspace and np.arange are built-in functions of the data mining library.…”

“…Huang Xiaogou et al [14] used the XGBoost algorithm to obtain damage identification for long-span bridge girders. Ling Dan et al [15] proposed the fault detection and identification method for the furnace negative pressure system based on canonical variable analysis (CVA) and XGBoost, which used CVA to reduce the data redundancy and construct regular residuals, and used the constructed typical residual variables to construct the fault monitoring model based on XGBoost. The damage diagnosis of cable-stayed bridges can be divided into two steps: damage location and damage quantification.…”

A survival prediction model based on PCA-HSIDA-LSSVM for patients with esophageal squamous cell carcinoma

“…Machine learning uses supervised learning, semi-supervised learning, unsupervised learning, reinforcement learning and ensemble learning to extract feature, cluster, classify and predict. 11–13 The survival prediction of ESCC patients is a prediction classification problem, including the processing of the dataset and the analysis of the relationships among the variables. 14,15 The prediction models based on machine learning has higher adaptability and accuracy than those based traditional applied statistics.…”