Diagnostic systems for turbocharged engines

A method for training an artificial neural network (ANN), the method comprising:

• determining a predicted value for an operating parameter measured by a first hardware sensor using data received from the ANN;
• determining an actual value for the operating parameter using data received from the first hardware sensor;
• and
• using the actual value to train the ANN if a difference between the actual and predicted values is less than a predetermined amount.

A machine-readable storage medium having stored thereon machine executable instructions, the execution of said instructions adapted to implement a method for diagnosing abnormal operation in a turbocharged engine having an engine control module (ECM) operable to control engine operation in response to data received from a plurality of sensors, the method comprising:

• storing in a memory data corresponding to:
• a compressor map defining a region of compressor efficiency and compressor speeds during operation;
• and
• a turbine map defining a region of turbine efficiency and turbine speeds during operation;
• determining a predicted value for an operating parameter measured by a first hardware sensor using data received from at least a second hardware sensor and the at least one of the compressor map and the turbine map;
• determining an actual value for the operating parameter using data received from the first hardware sensor;
• generating an abnormal operation signal if a difference between actual and predicted values is greater than a predetermined amount;
• disabling the first hardware sensor if the difference between the actual and predicted values has exceeded the predetermined amount a predetermined number of times or has occurred continuously over a predetermined period of time;
• and
• controlling operation of an engine based on the predicted value.

A machine-readable storage medium having stored thereon machine executable instructions, the execution of said instructions adapted to implement a method for training an artificial neural network (ANN), the method comprising:

• determining a predicted value for an operating parameter measured by a first hardware sensor using data received from the ANN;
• determining an actual value for the operating parameter using data received from the first hardware sensor;
• and
• using the actual value to train the ANN if a difference between the predicted and actual values is less than a predetermined amount.

An apparatus for diagnosing abnormal operation in a turbocharged engine having an engine control module (ECM) operable to control engine operation in response to data received from a plurality of sensors, the apparatus comprising:

• a microprocessor with a memory that includes data corresponding to:
• a compressor map defining a region of compressor efficiency and compressor speeds during operation;
• and
• a turbine map defining a region of turbine efficiency and turbine speeds during operation;
• a module configured to determine a predicted value for an operating parameter measured by a first hardware sensor using data received from at least a second hardware sensor and the at least one of the compressor map and the turbine map;
• a module configured to determine an actual value for the operating parameter using data received from the first hardware sensor;
• a module configured to generate an abnormal operation signal if a difference between actual and predicted values is greater than a predetermined amount;
• a module configured to disable the first hardware sensor if the difference between the actual and predicted values has exceeded the predetermined amount a predetermined number of times or has occurred continuously over a predetermined period of time;
• and
• a module configured to control operation of an engine based on the predicted value.

An apparatus for training an artificial neural network (ANN), the apparatus comprising:

• a module configured to determine a predicted value for an operating parameter measured by a first hardware sensor using data received from the ANN;
• a module configured to determine an actual value for the operating parameter using data received from the first hardware sensor;
• and
• a module configured to use the actual value to train the ANN if a difference between the predicted and actual values is less than a predetermined amount.