Fault Prognostic Based on Hybrid Method of State Judgment and Regression

Abstract

Fault prognostic is one of the most important problems in equipment health management system. This paper presents a hybrid method of mixture of Gaussian hidden Markov model (MG-HMM) and fixed size least squares support vector regression (FS-LSSVR) for fault prognostic. The system is established based on three parts. The first part trains the MG-HMM and FS-LSSVR model. According to the known samples, several MG-HMM models can be learned based on expectation maximization (EM) algorithm. Then, the forward variables can be calculated based on these MG-HMM models. Based on these forward variables, the corresponding FS-LSSVR models are built. All the MG-HMM models and corresponding FS-LSSVR models are combined into a model library. The second part recognizes the unknown sample based on the model library. This part obtains the MG-HMM model and FS-LSSVR model by maximization likelihood calculation between the unknown sample and MG-HMM models. The third part of the system calculates the forward variables based on the MG-HMM obtained from the second part. These forward variables are inputted into the corresponding FS-LSSVR model to compute the remaining useful life (RUL) of the unknown sample. Finally, we carry out experiments on benchmark data set to verify the proposed method. The results illustrate the effectiveness of the hybrid method.

Towards multi-model approaches to predictive maintenance: A systematic literature survey on diagnostics and prognostics

“…However, as mention in Section 5.2.3 these models still face problems as their results remain non-explicative [101]. Within the examples found in this literature review [159], presents a multi-layer perceptron and a radial basis function neural network in a parallel model to estimate the remaining useful life from input sensors on simulated jet-engines data [160]. performs fault prognostic with a mixture of Gaussian hidden Markov model (stochastic model) to evaluate the health index and fixed size least squares support vector regression (statistical model) for remaining useful life estimation on the same jetengines simulated data.…”

Real-Time Shop-Floor Production Performance Analysis Method for the Internet of Manufacturing Things

“…The extraction process from primitive event to basic event can be described by the pseudo code shown in Algorithm 1. / / represent the material number (4) if the object was first read by reader 1 at 1 //Reader 1 represent the entrance reader of the process line (5) then Create event BE 0,1 = ( .ID, 0 , 1(ts), launch) //New product launch (6) else if the object was last read by reader at 2 //Reader represent the exit reader of the process line (7) then Create event BE ,4 = ( .ID, , 2(te), finish) //The product is finished (8) else the object was read by reader at 3 //The reader in the process line (9) if first read (10) then Create event BE ,1 = ( .ID, , 3(ts), infield) //The infield event (11) and BE ,2 = ( .ID, −1 , 1(ts), 3(te), stay) //The stay event (12) and BE ,3 = ( .ID, −1 , 3(te), outfield) //The outfield event (13) Else Create event BE ,2 = ( .ID, −1 , 1(ts), 3( now), stay) //The stay event Algorithm 1: The extraction process from primitive event to basic event.…”

“…. ., attr , //Calculation based on the different complex event rule (9) Ts as ts, (10) Te as te (11) From CE, Calculation Process (12) Where CE.process.ID = CrE.process.ID (13) or CE.WIP.ID = part of product.ID Algorithm 3: The extraction process from complex event to critical event.…”

“…Zappia et al [12] proposed a lightweight and extensible complex event processing system based on a layered architectural design to extract meaningful events from raw data streams originated by sensing infrastructures. Li et al [13] presented a hybrid method of mixture of Gaussian hidden Markov model (MG-HMM) and fixed size least squares support vector regression (FS-LSSVR) for fault prognostic in equipment health management system. Xu and Liu [14] proposed a smart metering network system based on the IPv6 network protocol and ZigBee protocol for residential power measurement.…”

“…For the real-time manufacturing information processing, the event model has been widely adopted as seen in [9][10][11][12]. References [13][14][15][16] describe some predicted models for manufacturing exceptions. Obviously, the event model has advantages in dealing with the events which occurred at the Auto-ID devices.…”