| Method and system of monitoring and prognostics -> Monitor Keywords |
|
|
  Method and system of monitoring and prognosticsUSPTO Application #: 20070094173Title: Method and system of monitoring and prognostics Abstract: A neural network learns the operating modes of a system being monitored under normal operating conditions. Anomalies can be automatically detected and learned. A control command can be issued or an alert can be issued in response thereto. (end of abstract) Agent: Welsh & Katz, Ltd - Chicago, IL, US Inventors: Gregory A. Harrison, Alexey Turischev USPTO Applicaton #: 20070094173 - Class: 706016000 (USPTO) Related Patent Categories: Data Processing: Artificial Intelligence, Neural Network, Learning Task The Patent Description & Claims data below is from USPTO Patent Application 20070094173. Brief Patent Description - Full Patent Description - Patent Application Claims
FIELD OF THE INVENTION [0001] The invention pertains to detection of anomalies in complex circumstances and systems. More particularly, the invention pertains to the automatic detection of such anomalies while a system is being monitored. BACKGROUND OF THE INVENTION [0002] Neural network based analysis systems for detecting and analyzing vibration are known. One such system and method are disclosed in U.S. Pat. No. 6,301,572 B1 entitled "Neural Network Based Analysis Systems for Vibrational Analysis and Condition Monitoring" which was filed Oct. 9, 2001 and is assigned to the assignee hereof. The disclosure of the '572 patent is hereby incorporated herein by reference. [0003] In the system and method of the '572 patent, time domain outputs from a vibration sensor coupled to an apparatus being monitored are transferred to the frequency domain. Frequency domain outputs can then be provided as inputs to a fuzzy adaptive resonance-type neural network. Outputs from the network can be coupled to an expert system for analysis, to display devices for presentation to an operator or for use for other control and information purposes. [0004] While the system and method of the '572 patent are useful and effective for their intended purpose, that solution was directed primarily to addressing vibration signals. There is a need for and it would be desirable to be able to automatically detect anomalies in complex systems which are continually being monitored for any deviation from normal operating condition. It would be desirable if the monitoring system could automatically learn the characteristics of the anomalous condition and respond thereto by generating a control command or causing a selected indication to be produced. BRIEF DESCRIPTION OF THE DRAWINGS [0005] FIG. 1 is a block diagram of a system in accordance with the invention; [0006] FIG. 2 is a block diagram of a particular embodiment of a system as in FIG. 1; [0007] FIG. 3 is a block diagram of an embodiment of the invention which incorporates a plurality of neural networks; [0008] FIG. 4 is a block diagram of a different embodiment of the invention which incorporates a plurality of neural networking; and [0009] FIG. 5 is a block diagram of another embodiment in accordance with the invention. DETAILED DESCRIPTION OF THE EMBODIMENTS [0010] While embodiments of this invention can take many different forms, specific embodiments thereof are shown in the drawings and will be described herein in detail with the understanding that the present disclosure is to be considered as an exemplification of the principles of the invention and is not intended to limit the invention to the specific embodiment illustrated. [0011] In a system which embodies the invention, behaviors of various parameters which characterize an apparatus being monitored can be learned. Parameters include, an electrical state vector, the states of digital inputs and outputs, and operating modes of the apparatus. The system can then detect changes or variations from the learned behavior of those parameters. [0012] The detected changes or variations can be learned. They can trigger responses such as control command, alerts, recording of apparatus behavior or the like all without limitation. [0013] In one aspect of the invention, a system employs at least one neural network to learn the time-and mode-based characteristics of an external apparatus. Many interfaces are capable of being monitored, and actions may be taken through various outputs. The neural network learns the normal operation of the monitored apparatus. It can detect anomalies, or be used as a control system for the monitored apparatus. The system can serve as an embedded intelligence, with learning and acting capabilities, that can be applied in many situations. [0014] In another aspect of the invention, as the monitored apparatus is operated and changes modes, a neural network memory can be modified to record the traces of the system state as evidenced by the internal representation of an input signal. As the monitored apparatus is operated more and more, the traces in neural network memory get modified to form a general path through the recorded state space of the monitored apparatus. If the apparatus should deviate from ordinary operations then the deviation can be immediately detected. [0015] A neural network of a type disclosed, in U.S. Pat. No. 6,301,572 can be used to record the plurality of inputs presented to the network, during a learning phase. It can also be used to detect outlier input states during operation to produce an alert. [0016] The outlier inputs can represent an anomaly. At the least they represent a set of input states that have not been previously recorded. The anomaly path through state space/neural memory can be recorded in a neural net memory for later analysis, or to become part of the accepted state space path. [0017] The set of input states can enter the neural network through a neural dimension mapper that ensures that the inputs are correctly represented in a selected range in a neural interface vector. The neural dimension mapper can repeatedly create machine-learnable vectors of inputs that are passed to the neural network through a neural interface. [0018] In another aspect of the invention, the sensory input data can be obtained from a system that gathers the electrical sensor data and standardizes it to numerical ranges appropriate for the neural net. A hardware interface can be used to bring many different types of sensor input information to the neural net. [0019] In yet another aspect of the invention, a plurality of sensors can be coupled to a plurality of neural networks. The neural networks can each receive a common set of inputs. In accordance herewith, the sensors and neural networks combinations can be displaced from one another in wireless communication with, in one disclosed embodiment, a common control unit. [0020] In accordance with another disclosed embodiment, different sensors can provide inputs to different networks. A common device can be monitored or controlled by such a sensor/network combination. Continue reading... Full patent description for Method and system of monitoring and prognostics Brief Patent Description - Full Patent Description - Patent Application Claims Click on the above for other options relating to this Method and system of monitoring and prognostics patent application. ###  How KEYWORD MONITOR works... a FREE service from FreshPatents1. Sign up (takes 30 seconds). 2. Fill in the keywords to be monitored. 3. Each week you receive an email with patent applications related to your keywords. Start now! - Receive info on patent apps like Method and system of monitoring and prognostics or other areas of interest. ### Previous Patent Application: Spread kernel support vector machine Next Patent Application: Neural network based rating system Industry Class: Data processing: artificial intelligence ### FreshPatents.com Support Thank you for viewing the Method and system of monitoring and prognostics patent info. IP-related news and info Results in 0.18209 seconds Other interesting Feshpatents.com categories: Qualcomm , Schering-Plough , Schlumberger , Seagate , Siemens , Texas Instruments , |
||
