top of page
  • Writer's picture.

But we are running at 120%..

Updated: Apr 3, 2021

Leveraging Analytics for manufacturing performance

In asset intensive process industries, it is quite common to encounter this statement about higher than nameplate capacity utilisation and in that situation looking for opportunities for further improvement will be seen as an absolute futile exercise.

Whether technically 120% is possible or not is a matter of technical debate but there are some challenges in establishing the true capacities and current operating performance of these plants which lead to divergent views on this subject even between the individuals running such plants.

There are a few reasons for such difference of opinion

a)   Year on year additions in equipment and piping capacities have kept on enhancing overall capacity of the specific units which also have had a positive impact on the overall plant capacity but the nameplate capacities are not adjusted accordingly.

b)   Feedstock/Input material quality: variability of main feedstock and its technical composition (for example gross calorific value) can change the plant throughput by 10-20% leading to added difficulties in establishing the true plant potential.

c)    Capacity Creeps: In addition, the option of using substitute materials also create performance uplift but very rarely these change the nameplate capacity of the plant.

All of the above are tacit knowledge residing with long timers of the organisation and one may find it difficult to use these to establish the true potential. Despite these challenges, individuals & organisations looking to enhance utilisation levels can use some of the pointers below

1. Performance Analysis: Using deep data analytics, insights can be gathered on plant performance metrics on ‘favourable’ (higher quality input, ambient condition etc) and ‘unfavourable’ days. The key question to ask is given all favourable conditions, what parameters are leading to capacity loss or process variance. This knowledge can become very valuable to the organisation and the best way to gather is to involve operating team starting with the field & panel operators in this exercise.

 2. Setting up formal Capacity Loss Measurement Framework: While it may sound as the obvious thing to have and something existing for ages in companies, being ‘precise’ in allocating the losses needs expertise, collaboration and openness between functions. Especially in process manufacturing, allocating the losses in right manner can really help in building knowledge repository on the deviation & conditions leading to lower performance

 3. Alarms Data; Rightly set, the DCS systems and Advanced Process Controls systems generate tons of data on ‘deviations. Deeper analysis of alarms & incidents will also reveal the systems & equipment limitations which are holding the performance levels. 


4.Utilities Systems: The “Nerve Centre” of process plants are the utility systems. Making sure that they are running at their top efficiencies can ensure higher throughput & optimal energy efficiency. But it’s not uncommon see utilities running at lower levels and lack of organisational focus on the function.


5.Process Parameters Limits: Key parameters and their control limits are another critical lever which ensures process stability. Obvious but often not a common knowledge across the operating team with multiple difference of opinion on the exact control limits. Instead of going by past experiences and design parameters which may have changed due to capacity additions, it is advisable to use advanced analytics concepts to arrive at the most optimal parameters. While using advanced analytics, care is required to define the “objective function” and “control/limit parameters” in the most precise manner and it will be prudent to involve the operating team in defining these. Needless to say, use of analytics will only bring further refinement to the existing parameter set but the real challenge will always be on maintaining the plant with those parameter levels.


Lastly before trying Machine Learning to control deviations, it will be wise to ‘Learn more about the machines’. Equipment performance & reliability can explain majority of the deviations happening in process plants. Linking the DCS alarms with equipment’s performance and eliminating the issues on a systematic basis can help identify levers for taking the plant consistently to 120% utilisation levels & beyond.

36 views0 comments

Recent Posts

See All


bottom of page