Remaining useful life may be estimated for a machine component by training a prediction model, even when limited data from actual failures is available. Feature data such as sensor readings associated with a mechanical process may be collected over time. Such readings may be paired with estimates of remaining useful life, for instance as extracted from unstructured text of maintenance records. Such data may be used to train and test the prediction model.

An electronic device and a method are disclosed. The electronic device includes a sensor, a memory, a processor, and a communication interface. The sensor is configured to detected vibrations of a burner system including any component of a burner system. The memory is configured to store the detected vibrations. The processor is configured to record the detected vibrations caused by the burner system at a predetermined time interval. The processor is also configured to generate a report of the recorded vibrations caused by a burner component to indicate the operational status of the burner, wherein the generated report includes at least two recorded vibrations. The communication interface configured to transmit the generated report.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, that can adjust operations of a manufacturing plant based on an assessment of risk to the plant's operations posed by the conditions and/or operations of the different devices in the manufacturing plant. Methods may include obtaining, using a set of sensors, a set of current operational characteristics for a plurality of plant devices in a manufacturing plant. For a particular plant device, a set of risk factors corresponding to a failure of the particular plant device can be analyzed. Based on the set of risk factors, an overall risk posed by the particular plant device to operations of the manufacturing plant can be determined. Based on the overall risk, one or more operations of the manufacturing plant can be adjusted.

A quality control device that controls quality of a product manufactured through a plurality of processes, includes a prediction model generation unit that generates a prediction model to predict quality of a product with respect to unknown process data by performing learning using known process data obtained from the plurality of processes and a measured value of quality of the product with respect to the known process data as learning data; a quality prediction unit that derives a predictive value of quality of each of a plurality of products, which are manufactured after the prediction model is generated, on the basis of the prediction model using process data of the plurality of products as input data; and an inspection target decision unit that decides the product for which the predictive value having the smallest margin with respect to a preset standard is obtained as an inspection target, among the plurality of predictive values of quality obtained by the quality prediction unit.

A sensor management system includes a historical data repository, a building management system (BMS) controller, and a sensor diagnostics system. The historical data repository may store historical data by a plurality of sensors. The BMS controller may control one or more components of a building subsystem based on data provided by one or more sensors. The sensor diagnostic system may receive, from a sensor, sensor data. The sensor diagnostic system may determine at least one fault in the sensor data. The sensor diagnostic system may select, from the historical data repository, substitute sensor data for the sensor based on a comparison of one or more attributes of the sensor data and one or more attributes of the historical data in the data repository. The sensor diagnostic system may provide, in replacement of the sensor data from the sensor, the supplemental sensor data to the BMS controller.

An example method comprises receiving historical wind turbine failure data and asset data from SCADA systems, receiving first historical sensor data, determining healthy assets of the renewable energy assets by comparing signals to known healthy operating signals, training at least one machine learning model to indicate assets that may potentially fail and to a second set of assets that are operating within a healthy threshold, receiving first current sensor data of a second time period, applying a machine learning model to the current sensor data to generate a first failure prediction a failure and generate a list of assets that are operating within a healthy threshold, comparing the first failure prediction to a trigger criteria, generating and transmitting a first alert if comparing the first failure prediction to the trigger criteria indicates a failure prediction, and updating a list of assets to perform surveillance if within a healthy threshold.

An HVAC system includes a heating element, a discharge air temperature sensor, and a return air temperature sensor. A controller of the HVAC system determines that the HVAC system has been operating in the heating mode for at least a predefined amount of time. The controller receives measurements of the discharge air temperature and the return air temperature. A temperature rise value is determined using the discharge air temperature and return air temperature. If the temperature rise value is less than a predefined minimum threshold value, the controller determines that a first fault of the HVAC system is detected and provides a corresponding alert. If the temperature rise value is greater than a predefined maximum threshold value, the controller determines that a second fault of the HVAC system is detected and provides a corresponding alert.

Methods and systems automatically detect, classify and/or mitigate sensor errors using partial qualitative and quantitative knowledge of the subsystems. In various examples, sensor fault detection is performed with a custom designed representation scheme covering causality, correlation, system of equality and inequalities, and an associated logic. The logic is described by a set of algorithmic steps to iteratively assign trustworthiness level of sensors. Sensor fault classification is performed by combining mathematical and statistical techniques that can be utilized to expose bias, drift, multiplicative calibration error, precision degradation and spike error. Sensor fault mitigation is also performed on identified bias, drift, multiplicative calibration error, precision degradation and spike error.

A state monitor for monitoring the state of a system and including a calculator and a memory; the system being controlled by a command defining a plurality of modes of operation of the system, each mode of operation corresponding to applying a command of constant value; the memory containing a set of stored state matrices representing, for each mode of operation of the system, the value of the projection of its state in time; and the calculator being configured, during operation of the system, to determine estimated values for the state of the system at a given instant with the state functions and of its state at an earlier instant.

A valve abnormality detecting device includes an opening acquiring portion to acquire a valve opening value; a pressure acquiring portion to acquire a pressure value of operating device air of an operating device for a valve; a stability-time detecting portion configured to detect a stable-opening state in which the valve opening value acquired by the opening acquiring portion 1 is substantially constant; a frictional force detecting portion configured to detect a difference between a maximum pressure value and a minimum pressure value of the operating device air in the stable-opening state as an index value indicating a frictional force at a sliding portion of the valve; and an abnormality determining portion configured to determine that an abnormality may have occurred in the valve in a case where a frequency of occurrence of reduction in which the index value falls below a specified value is an abnormal frequency.