Product Details Supplier Info More products

CSense has enabled Sasol to identify the root cases of problems relating to production losses and the reduced efficiency levels of wastewater treatment at its water recovery plant.

These problems were being caused by breakthroughs of undesirable organics at the Sasol synfuels facility in Secunda.

Historically, a breakthrough of these organics only became evident a few days after its occurrence.

The organics cause the inhibition of the micro-organisms in the bio basins at the water recovery plant, resulting in unwanted foaming.

An anti-foaming agent then has to be dosed at a very high cost to Sasol.

The water recovery plant at the Sasol synfuels facility treats wastewater to produce cooling water for the factory.

Wastewater effluents consist of various streams that continuously vary in volume and composition.

The cooling water is processed from three sources: stripped gas liquor (SGL) from the phenosolvan plant, reaction water from chemical workup and storm and oily water rundown (API).

Since a variety of factors influence the performance of the water recovery plant, troubleshooting is very difficult.

Aerobic micro-organisms remove most of the contaminants from these sources in the bio basins at the plant.

These micro-organisms require optimum environmental conditions to efficiently upgrade the wastewater.

When environmental conditions such as pH, temperature, available food, available oxygen and the concentration of organics become undesirable, the micro-organisms will become dormant – a state that is characterised by severe foaming on the surfaces of the bio basins.

The foaming can be excessive to the extent that anti-foaming chemicals have to be sprayed onto the bio basins to avoid polluting the surrounding areas.

Anti-foaming dosage, in turn, carries a high operating cost for water recovery.

However, occurrences of foaming in the bio basins are indicative of a much higher production cost in that breakthroughs of the organic also cause upstream instabilities and process cooling water of varying quality.

As a result of inefficient treatment in the bio basins, sub-optimal process cooling water goes into the factory, increasing the fouling in heat exchangers and resulting in a further notable reduction in heat transfer and corresponding process efficiencies.

The latter factors represent the higher cost to the company and is difficult to quantify.

However, since efficiency was linked to the minimal use of the anti-foaming agent, the reduction in anti-foaming agent consumption was specified as the project goal.

Martin Hamann, senior process engineer at Sasol, saw that it was necessary to track these foaming incidents back through the process, considering many operating factors upstream at different production stages and identifying which operating conditions needed to be avoided in order to reduce the number of organic breakthroughs.

He said: ‘After years of confronting this persistent problem, it’s fair to say that we were more than a little surprised when CSense showed us what to do within days of implementation.

‘We followed the following process troubleshooting methodology with excellent results,’ added Hamann.

Data were selected from factors that were suspected to be responsible for the organic breakthroughs, but that could never be proved.

There data were integrated from multiple sources, which were a combination of manually captured data contained in log sheets and process data stored in historian databases.

CSense automatically merges the data from multiple sources, thus establishing a single data set for the troubleshooting exercise.

Two years’ worth of data was evaluated at six-hour intervals.

Real-time factors that were chosen included ambient temperature, tar separator levels at the tar separation plant and the pH of SGL.

Regarding the phenosolvan plant, real-time measurements were made of the temperature of the gas liquor (GL) sent to the plant, GL storage tank levels, treatment chemical ratios and phenosolvan loads.

Daily measurement factors that were chosen included SGL, API and reaction water chemical analyses as well as the nutrient dosing at the water recovery plant.

Bad quality data was removed by making use of the interactive and visual data preparation tools supplied by CSense.

It was said to be easy to estimate process lags between the foaming incidents and the deviation of upstream operating parameters.

These lags varied from hours to days.

The rules-extraction capability of CSense was then used to demarcate the multivariate operating conditions that needed to be prevented.

By implementing these process rules, the number of breakthroughs of the organic could be reduced significantly.

The results can now be used to justify upstream operational changes and projects to ensure that the operating conditions responsible for organic breakthroughs are prevented.

The most important conditions responsible for high anti-foaming agent consumption were determined to be: high GL temperatures, low GL storage tank levels and high tar separator levels; API contamination, high treatment chemical ratios at the phenosolvan plant and high tar separator levels; and API contamination, high treatment chemical ratios at the phenosolvan plant, high tar separator levels and insufficient nutrient dosing at water recovery.

Dealing with factual reality, rather than guesswork, has resulted in some important benefits for the Sasol synfuels facility.

These include reduced operating costs; it is estimated that the use of anti-foaming agent will be halved, thereby realising annual savings of around USD1m (GBP617,200).

Another benefit is improved water quality; consistently high-quality process water throughout the plant will lead to more effective heat transfer and corresponding process efficiencies.

This could represent significant savings.

Control variables that were previously thought to be insignificant or unrelated to the cause of the incidents have been identified as significant contributors to these process upsets.

Process lag times are now known and predictions for severe foaming can be made.

From a better understanding of the entire process, there is a projected reduction in upstream production downtime.

The rules for the service-level agreement between upstream process units (such as GL separation) and the phenosolvan and water recovery plants have now been established.

While the general behaviour of chemical processes is understood, understanding variances in that behaviour is almost never obvious.

With multiple interdependent variables at play in a real-time and hazardous environment, there is little room for solutions by successive approximation or guesswork.

Hamann said: ‘CSense proved to be successful in rapidly identifying the root causes of problems in a complicated multi-variant environment.

‘It is quick to implement and simple to understand and has proved accurate in calculating lag times while being able to work with both real-time and daily measurements,’ he added.

View full profile