Single-loop performance assessment
Score each closed loop against the best variance it could achieve, quickly and while it keeps running in automatic.
Find the loopsholding yourplant back.
Octograph assesses closed-loop control performance from the operating data your plant already records. No open-loop tests, no switching controllers off, no process model.
What Octograph does
Octograph looks across the control loops in a processing plant and shows where control is costing you production, then what to change.
Focus where performance is slipping
Octograph ranks loops by how far their performance has degraded, so improvement work goes where it returns the most plant performance.
Optimal control coupling
Octograph works out which process variable each manipulated variable should control to increase production output.
How an assessment runs
From data your historian already holds to a ranked list of loops and pairings, without disturbing the plant.
Read routine data
Octograph works from closed-loop operating data. Controllers stay in automatic and there are no bump tests.
Estimate dead time
Each loop's process dead time is estimated online, in the presence of noise, rather than assumed as a constant.
Benchmark each loop
Loop variance is compared with the minimum-variance benchmark of the Harris Index, using the estimated dead time.
Rank the plant
The assessment extends plant-wide without needing process or control system knowledge, and loops are ranked by improvement potential.
Recommend pairings
Octograph identifies how process and manipulated variables should be paired to lift production.
Research behind Octograph
Peer-reviewed papers presented at Asian, Australian, and Australian and New Zealand control conferences between 2015 and 2017.
-
2015
On the industrial plant performance & operating point drifting phenomenon
10th Asian Control Conference (ASCC)
Examines what drives effective performance in processing plants, from equipment availability and operating practice to consistent, high-performing control. Coal handling preparation plants and bauxite beneficiation plants are assessed against these factors, covering both the process and the control schemes that support it.
View on IEEE Xplore -
2016
Extending the Harris Index performance assessment technique: a plant-wide focus
Australian Control Conference (AuCC)
Proposes two extensions to the Harris Index. The first improves accuracy by extracting process dead time from closed-loop data instead of relying on generalised prediction-horizon constants. The second scales the assessment to a whole plant without requiring process or control system knowledge.
View on IEEE Xplore -
2017
An on-line process dead-time estimation algorithm
11th Asian Control Conference (ASCC)
An algorithm that estimates process dead time within a closed-loop, time-invariant system, without open-loop tests or disabling the controller, and with noise present in the loop. The estimate can replace the constant dead-time assumption in assessments such as the Harris Index.
View on IEEE Xplore -
2017
A methodology to determine the dynamic relationship between process and manipulated variables
Australian and New Zealand Control Conference (ANZCC)
A method for identifying how pairing process and manipulated variables improves performance, including whether a loop exists, whether it runs in manual or automatic, and which coupling maximises production. An industrial demonstration shows the improvement available.
View on IEEE Xplore