Cascaded control of an average value of a process parameter to a desired value

A controller for directing operation of a system performing a process, the process having multiple process parameters (MPPs), including at least one controllable process parameter (CTPP) and at least one targeted process parameter (TPP), and having a defined target value (DTV) representing a first limit on an actual average value (MV) of the TPP over a time period (TP), comprising:

• one of a neural network process model and a non-neural network process model, the one model representing a relationship between the TPP and the at least one CTPP;
• first logic to predict a path corresponding to future average values (FAVs) of the TPP over a first time period (FTP) extending from a current time T0 to a future time TF1, prior to which the TPP will move to a steady state condition, and having a length of at least TP, based on (i) the AAVs of the TPP at various times over a first prior time period having a length of at least TP and extending from a prior time T−F1 to the current time T0, (ii) the current MPPs, (iii) the DTV, and (iv) the one model;
• and
• second logic to establish a further target value (FTV) representing a second limit on the AAV of the TPP for a second time period (STP) extending from the current time T0 to a future time TF2 and having a length less than the FTP, based on the predicted path, to determine a target set point for each CTPP based on the FTV and the one model, and to direct control of the system operations based on the target set point for each CTPP.

A method for directing performance of a process, having multiple process parameters (MPPs), at least one of the MPPs being a controllable process parameter (CTPP) and one of the MPPs being a targeted process parameter (TPP), and having a defined target value (DTV) representing a first limit on an actual average value (AAV) of the TPP over a defined time period of length TPLAAV2, with the MV computed based on actual values (AVs) of the TPP over the defined period, comprising:

• predicting future average values (FAVs) of the TPP over a first future time period (FFTP) having a length of at least TPLAAV2 and extending from a current time T0 to an future time TAAV2, at or prior to which the TPP will move to steady state, wherein the FAVs are predicted based on (i) the AAVs of the TPP at various times over a first prior time period (FPTP) having a length of at least TPLAAV2 and extending from a prior time of T−AAV2 to the current time T0, (ii) the current values of the MPPs, and (iii) the DTV;
• establishing a further target value (FTV) representing a second limit on the AAV of the TPP at the end of a second future time period (SFTP), the SFTP having a length equal to TPLAAV1 which is less than the length TPLAAV2 and extending from the current time T0 to a future time TAAv1, wherein the FTV is established based on one or more of the predicted FAVs of the TPP over the FFTP;
• determining a target set point for each CTPP based on (i) the AAVs of the TPP at various times over a second prior time period (SPTP) having the length TPLAAV1 and extending from a prior time T−AAV1 to the current time T0, (ii) the current values of the MPPs, and (iii) the FTV;
• directing control of each CTPP in accordance with the determined target set point for that CTPP.

A method for directing control of the performance of a process, the process having multiple process parameters (MPPs), including at least one controllable process parameter (CTPP) and at least one targeted process parameter (TPP), and having a defined target value (DTV) representing a first limit on an actual average value (AAV) of the TPP over a time period (TP), comprising:

• predicting a path corresponding to future average values (FAVs) of the TPP over a first time period (FTP) extending from a current time T0 to a future time TF1, prior to which the TPP will move to a steady state condition, and having a length of at least TP, based on (i) the AAVs of the TPP at various times over a first prior time period having a length of at least TP and extending from a prior time T−F1 to the current time T0, (ii) the current MPPs, (iii) the DTV, and (iv) one of a neural network model and a non-neural network model, the one model representing a relationship between the TPP and the at least one CTPP;
• establishing a further target value (FTV) representing a second limit on the AAV of the TPP for a second time period (STP) extending from a current time T0 to a future time TF2 and having a length less than the FTP, based on the predicted path;
• determining a target set point for each CTPP based on the established FTV and the one model;
• and
• directing control of performance of the process based on the target set point for each CTPP.