FAILURE DETECTION OF SAMPLE INTRODUCTION SYSTEMS
A method of operating a sample introduction system of an inductively coupled plasma analytical instrument, the method comprising applying a trained classifier to instrument data, obtained from the analytical instrument, during operation of the analytical instrument, to detect whether the sample introduction system is operating in a normal state or in a failure state. The method further comprises activating an error procedure in the event that the sample introduction system is operating in a failure state. The instrument data comprises signal data obtained from an analytical measurement made by the analytical instrument. The trained classifier is trained using a training data set comprising instrument data corresponding to the normal state of the sample introduction system.
1 . A method of operating a sample introduction system of an inductively coupled plasma analytical instrument, the method comprising:
applying a trained classifier to instrument data, obtained from the analytical instrument, during operation of the analytical instrument, to detect in which operating state of a plurality of operating states the sample introduction system is operating, wherein the plurality of operating states includes a normal state and a failure state; and
activating an error procedure in response to detecting that the sample introduction system is operating in a failure state;
wherein the instrument data comprises signal data obtained from an analytical measurement made by the analytical instrument; and
wherein the trained classifier is trained using a training data set comprising instrument data corresponding to the normal state of the sample introduction system.
2 . The method of claim 1 , wherein the failure state includes a plurality of failure sub-states each corresponding to one of a plurality of failure categories.
3 . The method of claim 1 , wherein the plurality of operating states further includes a close-to-failure state.
4 . The method of claim 3 , wherein the close-to-failure state includes a plurality of close-to-failure sub-states each corresponding to one of a plurality of failure categories.
5 . The method of claim 2 , wherein the plurality of failure categories comprises at least one of:
a leaking component of the sample introduction system;
a clogged component of the sample introduction system;
a damaged component of the sample introduction system; and
a flow through a component of the sample introduction system that deviates from an expected flow.
6 . The method of claim 5 , wherein the plurality of failure sub-states comprises at least one of:
a leaking sample tube;
a clogged nebulizer;
a nebulizer flow that deviates from what an expected nebulizer flow;
a leaking peristaltic pump tube;
a damaged peristaltic pump tube; and
an empty sample vial.
7 . The method of claim 1 , wherein the sample introduction system comprises one or more sensors, and wherein the instrument data further comprises sensor data comprising outputs from the one or more sensors.
8 . The method of claim 1 wherein the signal data comprises data that is representative of a property of the inductively coupled plasma, wherein optionally the data that is representative of a property of the inductively coupled plasma comprises an amount of a first species present in the plasma.
9 . The method of claim 8 wherein the signal data comprises data that is a ratio of the amount of the first species present in the plasma and a second species present in the plasma, wherein optionally the signal data comprises a ratio of the amount of Argon in the plasma and the amount of nitrogen in the plasma.
10 . The method of claim 8 wherein data that is representative of an amount of a species in the plasma comprises a recorded intensity of species emissions in the plasma.
11 . The method of claim 1 wherein the signal data comprises spectrometric data.
12 . The method of claim 1 wherein the trained classifier comprises a trained machine learning algorithm, wherein optionally the trained machine learning algorithm comprises a neural network.
13 . The method of claim 1 wherein the training data set further comprises instrument data corresponding to the failure state of the sample introduction system.
14 . The method of claim 7 , wherein the sensor data comprises data obtained from at least one of:
a nebulizer backpressure sensor;
a nebulizer flow sensor;
a cooling gas flow sensor;
a radio frequency plasma power sensor; and
a peristaltic pump speed sensor.
15 . The method of claim 1 , wherein a first activation function of the trained classifier comprises a rectified linear function and wherein optionally a second activation function of the trained classifier comprises a softmax function.
16 . The method of claim 1 wherein a loss function of the trained classifier comprises a categorical cross entropy function.
17 . The method of claim 1 wherein the error procedure comprises at least one of:
notifying the failure state to a user; and
placing at least one of the sample introduction system the analytical instrument into a safe mode, wherein optionally the safe mode comprises any of:
stopping the sample introduction system;
stopping one or more components of the sample introduction system; and
preventing a sample from entering a nebulizer of the sample introduction system.
18 . The method of claim 1 further comprising generating the trained classifier by performing Adam optimization on an initial classifier using the training data set.
19 . An apparatus arranged to carry out a method according to claim 1 .
20 . A computer-readable medium storing a computer program which, when executed by a processor, causes the processor to carry out a method according to claim 1 .
21 . (canceled)