Hybrid partial and full step quadratic solver for model predictive control of diesel engine air path flow and methods of use
Abstract
Methods and systems for use of model predictive control (MPC) controllers utilizing hybrid, quadratic solvers to solve a linear feasibility problem corresponding to a nonlinear problem for an internal combustion engine plant such as a diesel engine air path. The MPC solves a convex, quadratic cost function having optimization variables and constraints and directs the plant per the output solutions to optimize plant operation while adhering to regulations and constraints. The problem includes a combination of iterative and direct calculations in the primal space depending on whether a partial step (iterative) or a full step (direct) is attempted. Further, primal and dual space array matrices are pre-computed and stored offline and are retrieved via use of a unique identifier associated with a specific active set for a set of constraints. Such hybrid and/or offline calculations allow for a reduction in computational power while still maintaining accuracy of solution results.
Claims
exact text as granted — not AI-modifiedWhat is claimed is:
1 . A method for controlling an internal combustion engine having a variable geometry turbine (VGT), an exhaust gas recirculation (EGR) valve, and an EGR throttle, the method comprising:
solving a linear quadratic problem with a predictive model comprising an updating algorithm in order to determine: (i) a requested optimized VGT lift that meets one or more constraints; and (ii) a requested optimized EGR valve flow rate that meets the one or more constraints, wherein solving the linear quadratic problem includes:
determining whether to take a determined step comprising one of a primal partial step and a primal full step at each iteration, and
taking the determined step at each iteration until the linear quadratic problem is solved by the updating algorithm, wherein:
taking the primal partial step comprises performing an iterative calculation, and
taking the primal full step comprises performing a direct calculation;
generating the requested optimized VGT lift responsive to an engine intake manifold pressure by controlling the VGT; and generating the requested optimized EGR valve flow rate responsive to an EGR rate by controlling the EGR valve and the EGR throttle.
2 . The method of claim 1 , further comprising:
setting one or more engine operating parameters as the one or more constraints to form a non-linear problem; and deriving the linear quadratic problem based on the non-linear problem, wherein the linear quadratic problem is convex and time-varying.
3 . The method of claim 2 , wherein the linear quadratic problem to solve comprises the following equation:
min
J
(
)
=
1
2
T
+
ℋ
T
s
.
t
.
≤
where is a time varying linear term in the cost function; is a constant quadratic term in the cost function; is a constant constraint matrix; is a time varying constraint vector; and is one or more optimization variables comprising time varying control inputs.
4 . The method of claim 3 , wherein the time varying control inputs comprise at least one of the a VGT lift and an EGR flow, and further comprising adding Lagrangian multipliers μ to determine a solution for optimization variable(s) * and Lagrangian multiplier(s) μ* that are greater than or equal to zero to satisfy dual feasibility, while utilizing only active, feasible constraints of the one or more constraints.
5 . The method of claim 4 , wherein:
the VGT comprises turbine input vanes configured to be angled to be opened, partially opened, or closed; and the active, feasible constraints comprises at least one of a maximum EGR rate, a minimum EGR rate, a maximum EGR flow command, a minimum EGR flow command, a maximum VGT lift closed command to control a maximum amount of closure of the turbine input vanes, and a minimum VGT lift closed command to control a minimum amount of closure of the turbine input vanes.
6 . The method of claim 5 , further comprising simplifying the determined solution for a current active set list L and to define iterative and direct calculation approaches to determine how to move from the current active set list L to a new active set list L+1, resulting in the following simplified HQPKWIK equation:
L
+
1
=
{
L
+
zv
knext
t
Primal
Partial
(
Half
)
Step
-
T
L
+
1
U
(
T
L
+
1
U
)
T
ℋ
+
T
L
+
1
C
R
L
+
1
-
T
L
+
1
Primal
Full
Step
μ
L
+
1
=
μ
L
-
rv
knext
t
Dual
Step
where t is a minimum length that still maintain dual feasibility, v knext is a next constraint to be added to a current constraint set list, z is an array that is indicative of a search direction in primal space, and r is an array that is indicative of a search direction in dual space, such that:
z=T L U ( T L U ) T
r=R L 7 −1 ( T L C ) T .
7 . The method of claim 6 , further comprising computing the z and r arrays on-line.
8 . The method of claim 6 , further comprising pre-computing portions of the z and r arrays off-line.
9 . The method of claim 8 , further comprising retrieving the z and r arrays after a determination of a unique identifier at least partially based on the current active set list.
10 . The method of claim 9 , wherein the determination of the unique identifier utilizes a conversion of one or more binary numbers of the current active set list to a decimal value, and the unique identifier comprises the decimal value.Join the waitlist — get patent alerts
Track US2019107072A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.