equinor/TimeSeriesAnalysis

View on GitHub

Library for dynamic time-series modeling, identification and simulation. Focus on dealing with non-ideal real world datasets, and applications to industrial processes and -pid-feedback loops. Robust and fast for advanced analytics. Built on .NET to run anywhere.

Stars: 22
Forks: 5
Open beginner issues: 0
Indexed issues: 53

Dominant language: C#
License: Apache-2.0
Last GitHub push: Jun 19, 2026
Latest indexed: Sep 15, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: No beginner labels indexed
PR merge metrics: No merged PRs in 30d

Open All

53 open issues indexed

All methods of Vec will be ported to the generic Vec<t> when upgrading to a version of .NET that supports generic math Open

enhancement

equinor/TimeSeriesAnalysis#23 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Add functionality for two time-constants acting on the same output Open

enhancement

equinor/TimeSeriesAnalysis#40 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
add a "sum" block Open

enhancement

equinor/TimeSeriesAnalysis#42 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
add example of bumpless transfer in/out of automatic/manual mode Open

enhancement simulation

equinor/TimeSeriesAnalysis#50 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
add example showing that anti-windup works Open

enhancement

equinor/TimeSeriesAnalysis#51 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
add in and write example for split range control Open

enhancement simulation

equinor/TimeSeriesAnalysis#59 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PlantSimulator should try simulating sub-models in the order they are given, if initalization or simulation fails for whatever reason -allowing the user a path to override internal logic for hard cases Open

enhancement simulation

equinor/TimeSeriesAnalysis#62 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Steady-state initalization of models "in transient" Open

enhancement simulation

equinor/TimeSeriesAnalysis#63 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Code to initalize select controllers is not completely general Open

simulation

equinor/TimeSeriesAnalysis#69 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
nonlinear models :consider if the choice of u0 has any impact Open

identification

equinor/TimeSeriesAnalysis#76 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
If non-white noise is added to an output it appears that the time constants are over-estimated(minor bug) Open

bug

equinor/TimeSeriesAnalysis#82 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Add a plot to the simulator which illustrates the flow diagram of how UnitIdentifier works internally Open

documentation

equinor/TimeSeriesAnalysis#83 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Add a plot which explains the logic of how PlantSimulator initializes simulations in steady-state Open

documentation

equinor/TimeSeriesAnalysis#84 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
As N increases, time-constant estimates suffer (noise enters model through diff-) Open

bug enhancement

equinor/TimeSeriesAnalysis#85 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PidIdentfier should give out uncertainty on parameters (like UnitIdentifier does) Open

equinor/TimeSeriesAnalysis#102 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
verify that identification is able to take the effect of specified scaling into account Open

equinor/TimeSeriesAnalysis#104 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
make it possible to adjust if sas scales kp from input object (so that it can be affected from front-end) Open

equinor/TimeSeriesAnalysis#105 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopUnitIdentifier only works for a single PID-controller Open

enhancement

equinor/TimeSeriesAnalysis#114 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
process identification does not work for processs controlled by gain-scheduling Open

bug known-limitation

equinor/TimeSeriesAnalysis#115 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Pid identification currently does not work for gain-scheduled systems Open

known-limitation

equinor/TimeSeriesAnalysis#116 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopIdentifier does not attempt to estimate time delays Open

ClosedLoopIdentifier known-limitation

equinor/TimeSeriesAnalysis#118 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
extrapolation for U outside Umin/Umax in tuning set should be only linear Open

enhancement

equinor/TimeSeriesAnalysis#119 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ProcessControl examples runs, but errounously has warning "PlantSimulatorInitializer: computational loop init failed" Open

bug

equinor/TimeSeriesAnalysis#123 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Use MinVer Open

enhancement

equinor/TimeSeriesAnalysis#125 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
better python integration (pythonnet?) Open

equinor/TimeSeriesAnalysis#133 · 1 comment · 0 reactions · 1 assignee · Indexed Sep 15, 2026
GainSchedIdentifier.IdentifyForGivenThresholds does not currently support identifying time-constants if Tc-threshold do not match gain-thresholds Open

Gain-Scheduling known-limitation

equinor/TimeSeriesAnalysis#137 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 15, 2026
GainSchedIdentifier.Identify will only consider suggesting a single threshold (could also consider multiple) Open

Gain-Scheduling known-limitation

equinor/TimeSeriesAnalysis#139 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Build is market as failing, beause .snupkg does not inlclude a *.pdb file Open

bug

equinor/TimeSeriesAnalysis#142 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
GainSchedIdentifier: ignoring given indices not implemented Open

Gain-Scheduling known-limitation

equinor/TimeSeriesAnalysis#145 · 1 comment · 0 reactions · 1 assignee · Indexed Sep 15, 2026
GainSchedIdentifier: experimental code to find two time-constants has been commented out of Identify() Open

enhancement Gain-Scheduling

equinor/TimeSeriesAnalysis#146 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
GainSchedIdentifier.Identify() uses two methods for selecting the best models among a candidate set Open

enhancement Gain-Scheduling

equinor/TimeSeriesAnalysis#147 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Make a unit test to investigate integral oscillations Open

enhancement PidIdentifier

equinor/TimeSeriesAnalysis#149 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
pid identification does not work well when a pid-controller is saturated in portions of the dataset Open

enhancement PidIdentifier

equinor/TimeSeriesAnalysis#150 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopUnitIdentifier does not take FittingSpecs specifications into account. Open

ClosedLoopIdentifier enhancement

equinor/TimeSeriesAnalysis#152 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Consider using splines in GainSchedModel, as a way to get "smoother" models without requiring a large number of thresholds. Open

enhancement Gain-Scheduling

equinor/TimeSeriesAnalysis#154 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PidIdentifier should have its own FittingSpecs Open

enhancement PidIdentifier

equinor/TimeSeriesAnalysis#158 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
if u is outside range [0,100] pididentifier now "fails quietly". Open

enhancement PidIdentifier

equinor/TimeSeriesAnalysis#161 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Identifier: Missing unit on uncertainty of identified time constant Open

documentation enhancement

equinor/TimeSeriesAnalysis#162 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PlantSimulator initalization does not account for saturated pid-controllers Open

known-limitation PlantSimulator

equinor/TimeSeriesAnalysis#165 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopIdentifier "random walk" disturbance tests work for some seeds but not for others Open

ClosedLoopIdentifier known-limitation

equinor/TimeSeriesAnalysis#167 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopIdentifier is not able to identify gains in cases that there are nonzero disturbances in a multiple-input system. Open

ClosedLoopIdentifier known-limitation

equinor/TimeSeriesAnalysis#168 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
ClosedLoopIdentifier struggles when a time-constant is introduced along with random walk disturbances. Open

ClosedLoopIdentifier known-limitation

equinor/TimeSeriesAnalysis#169 · 2 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Cannot import TimeSeriesDataSet class in Pythonnet Open

equinor/TimeSeriesAnalysis#173 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PidIdentifier should ideally rely only of FitScore to select among candidate models. Open

known-limitation

equinor/TimeSeriesAnalysis#195 · 1 comment · 0 reactions · 1 assignee · Indexed Sep 15, 2026
All identification method need to parse all timestamps for regularity before proceeding Open

bug enhancement

equinor/TimeSeriesAnalysis#196 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
UnitTimeDelayIdentifier should ideally only use FitScore to select models. Open

enhancement

equinor/TimeSeriesAnalysis#197 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PidIdentifier tends to under-estimate Kp and Ti by 20% on under-sampled datasets. Open

enhancement identification PidIdentifier

equinor/TimeSeriesAnalysis#208 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Plot class has hard-coded adresses and only works on PC (not mac) Open

known-limitation

equinor/TimeSeriesAnalysis#214 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
PlantSimulator does not appear to obey "Ymin" and "Ymax" in simulations of UnitModel Open

bug

equinor/TimeSeriesAnalysis#215 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
In some cases the UnitIdentifier will return Infinity gains if the input is flat and the output varies. Open

bug

equinor/TimeSeriesAnalysis#216 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026

equinor/TimeSeriesAnalysis

Get new issues in your inbox