Neural network and method for estimating regions of motor operation from information characterizing the motor
View Patent ↗A method for collecting operational parameters of a motor may include controlling the energization of a phase winding of the motor to establish an operating point, monitoring operational parameters of the motor that characterize a relationship between the energization control applied to the motor's phase winding and the motor's response to this control, and collecting information of the operational parameters for the operating point that characterizes the relationship between the applied energization control and the motor's response. The collected information characterizing the relationship between the applied energization control and the motor's response may be employed by a neural network to estimate the regions of operation of the motor. And a system for controlling the operation of motor may employ this information, the neural network, or both to regulate the energization of a motor's phase winding during a phase cycle.
1. A neural network for estimating the position of a motor's rotor with respect to its stator, the network comprising:
a preprocessor that produces a substantially non-linear response based on a plurality of variables; and
a plurality of neurons, wherein:
an input layer of the neural network comprises a group of the plurality of neurons and each neuron of the input layer produces a response based on the plurality of variables and the response produced by the preprocessor,
an output layer of the neural network comprises one of the plurality of neurons, which produces a response based on the responses produced by the neurons of the input layer, and
information indicative of: (1) the amount of current applied to a phase winding of the motor and (2) the motor's flux-linkage are variables provided to the preprocessor for use in producing the substantially non-linear response and also provided to each of the input layer neurons for use in producing its respective response.
2. The neural network of claim 1 , wherein the input layer comprises all but one of the neurons.
3. The neural network of claim 2 , wherein the input layer consists of three neurons.
4. The neural network of claim 1 , wherein the preprocessor produces the non-linear response by multiplying two of the variables together.
5. The neural network of claim 1 , wherein the preprocessor produces the non-linear response by dividing one of the variables by another.
6. The neural network of claim 1 , wherein the position is estimated for the rotor of a switched reluctance motor.
7. A neural network for estimating a value, the network comprising:
a preprocessor that produces a substantially non-linear response based on a plurality of variables; and
a plurality of neurons, wherein:
an input layer of the neural network comprises a group of the plurality of neurons and each neuron of the input layer produces a response based on the plurality of variables and the response produced by the preprocessor,
an output layer of the neural network comprises one of the plurality of neurons, which produces a response based on the responses produced by the neurons of the input layer, and
each of the variables provided to the preprocessor for use in producing the substantially non-linear response is also provided to each of the input layer neurons for use in producing its respective response.
8. The neural network of claim 7 , wherein the input layer comprises all but one of the neurons.
9. The neural network of claim 8 , wherein the input layer consists of three neurons.
10. The neural network of claim 7 , wherein the preprocessor produces the non-linear response by multiplying two of the variables together.
11. The neural network of claim 7 , wherein the preprocessor produces the non-linear response by dividing one of the variables by another.
12. A method for estimating the position of a motor's rotor with respect to its stator using a neural network, the method comprising:
generating a substantially non-linear response based on a plurality of variables;
generating, for each input layer neuron of the network, a response based on the plurality of variables and the substantially non-linear response; and
generating a network response based on the responses generated by the input layer neurons, wherein
information indicative of: (1) the amount of current applied to a phase winding of the motor and (2) the motor's flux-linkage are variables provided to the preprocessor for use in producing the substantially non-linear response and also provided to each of the input layer neurons for use in producing its respective response.
13. The method of claim 12 , wherein the substantially non-linear response is generated by multiplying two of the variables together.
14. The method of claim 12 , wherein the substantially non-linear response is generated by dividing one of the variables by another.
15. The method of claim 12 , wherein the neural network input layer consists of three neurons, a single output layer neuron produces the network response, and no neuron exists in an intermediate layer between the input layer and the output layer.
16. The method of claim 12 , wherein the position is estimated for the rotor of a switched reluctance motor.
17. A method for estimating a value using a neural network, the method comprising:
generating a substantially non-linear response based on a plurality of variables;
generating, for each input layer neuron of the network, a response based on the plurality of variables and the substantially non-linear response; and
generating a network response based on the responses generated by the input layer neurons, wherein
each of the plurality of variables provided to the preprocessor for use in producing the substantially non-linear response is also provided to each of the input layer neurons for use in producing its respective response.
18. The method of claim 17 , wherein the substantially non-linear response is generated by multiplying two of the variables together.
19. The method of claim 17 , wherein the substantially non-linear response is generated by dividing one of the variables by another.
20. The method of claim 17 , wherein the neural network input layer consists of three neurons, a single output layer neuron produces the network response, and no neuron exists in an intermediate layer between the input layer and the output layer.