System and method for retroactive and automated validation or corrective action with respect to online sensors
A system and method is provided for correction of sensors in an industrial process. A first sensor provides first data corresponding to measured variables for at least one respective process component, which are further aggregated in data storage with corresponding contextual tags comprising time series identifiers. For a given sample from the industrial process, at least the first data having a corresponding time series identifier are obtained for comparison with second data collected via a second sensor and which comprises measured variables corresponding to the at least one respective process component from the given sample. A feedback signal is selectively generated based on a determined error between the collected first data and the collected second data corresponding to the given sample, and may in an embodiment be used for automatically calibrating the first sensor. The feedback signal may further or alternatively be used to prompt interventions in the first sensor.
1 . A method for correction of sensors in an industrial process, the method comprising:
collecting first data via one or more online first sensors with respect to the industrial process, the first data corresponding to measured variables for at least one respective process component;
aggregating the first data in data storage as respective input data points with corresponding contextual tags comprising time series identifiers;
for a given sample collected from a production stage of the industrial process and having a respective time series identifier, obtaining at least one respective input data point of the first data having a corresponding time series identifier;
collecting second data via one or more second sensors, the second data comprising measured variables corresponding to the at least one respective process component from the given sample;
determining a difference between at least one measured variable from the first data and the corresponding at least one measured variable from the second data;
selectively generating a feedback signal to at least one of the one or more first sensors as an error value based on the determined difference; and
automatically calibrating the at least one of the one or more first sensors to correct further first data collected from the at least one of the one or more first sensors based on the generated feedback signal.
2 . The method of claim 1 , further comprising:
automatically determining an intervention state for at least one of the one or more first sensors based at least in part on the feedback signal and generating a message corresponding to the determined intervention state via a user interface.
3 . The method of claim 2 , wherein at least one type of determined intervention state corresponds to a message prompting maintenance of the at least one of the one or more first sensors.
4 . The method of claim 2 , wherein at least one type of determined intervention state corresponds to a message prompting replacement of the at least one of the one or more first sensors.
5 . The method of claim 2 , wherein the intervention state is determined based on predictive analysis with respect to an error condition.
6 . The method of claim 5 , wherein the intervention state is further determined in view of a margin for error and utilizing observed trends corresponding to the aggregated first data.
7 . The method of claim 6 , wherein the margin for error is dynamically determined based on the observed trends corresponding to the aggregated first data.
8 . The method of claim 1 , further comprising:
adjusting the aggregated first data based on the feedback signal to retroactively account for the determined error value, further in view of the corresponding time series identifier.
9 . The method of claim 1 , wherein at least the determined error value is implemented for development and/or training of a model to reflect learned relationships between the first data and one or more process conditions and/or quality metrics.
10 . The method of claim 1 , wherein the one or more second sensors are offline with respect to the industrial process.
11 . The method of claim 1 , wherein the one or more first sensors are mounted at a first online location to generate first data for the given sample with a first time series identifier, and the one or more second sensors are mounted at a second online location to generate second data for the given sample with a second time series identifier.
12 . The method of claim 1 , wherein the one or more first sensors are mounted at a first online location to generate first data for the given sample with a first time series identifier, the one or more second sensors are mounted at a second online location to generate data for the given sample with a second time series identifier, and the data generated by the one or more second sensors are processed to indirectly determine the second data corresponding to the at least one respective process component.
13 . A system comprising:
one or more first sensors mounted online with respect to an industrial process, and configured to generate first data corresponding to measured variables in real time for at least one respective process component;
data storage aggregating the first data as respective input data points with corresponding contextual tags comprising time series identifiers; and
a data processor configured, for a given sample collected from a production stage of the industrial process and having a respective time series identifier, to
obtain at least the respective input data point of the first data having a corresponding time series identifier,
collect second data via one or more second sensors, the second data comprising measured variables corresponding to the at least one respective process component from the given sample,
determine a difference between at least one measured variable from the first data and the corresponding at least one measured variable from the second data, and
selectively generate a feedback signal to at least one of the one or more first sensors as an error value based on the determined difference;
wherein a controller is functionally linked to the at least one of the one or more first sensors and configured to automatically correct further first data generated from the at least one of the one or more first sensors based on the generated feedback signal.
14 . The system of claim 13 , wherein the processor is further configured to adjust the aggregated first data based on the feedback signal to retroactively account for the determined error value, further in view of the corresponding time series identifier.
15 . The system of claim 13 , wherein the processor is further configured to implement at least the determined error value for development and/or training of a model to reflect learned relationships between the first data and one or more process conditions and/or quality metrics.
16 . The system of claim 13 , wherein the one or more second sensors are offline with respect to the industrial process.
17 . The system of claim 13 , wherein the one or more first sensors are mounted at a first online location to generate first data for the given sample with a first time series identifier, and the one or more second sensors are mounted at a second online location to generate second data for the given sample with a second time series identifier.
18 . The system of claim 13 , wherein the one or more first sensors are mounted at a first online location to generate first data for the given sample with a first time series identifier, the one or more second sensors are mounted at a second online location to generate data for the given sample with a second time series identifier, and the data generated by the one or more second sensors are processed to indirectly determine the second data corresponding to the at least one respective process component.