Concentration or dilution control: Choosing a control approach
Some of the simplest processes are dilution or mixing, that is, the combining together of products; separation, the removal of one or more products from a mix or blend and reaction monitoring where the products react together to form a new product.
All of these operations require control. Control requires some form of instrumentation and control equipment.
In any process, the choice of method and instrumentation depends on the process objectives and the fluids involved. For example, in caustic blending, we may not measure concentration directly but we can measure other fluid properties that change with concentration, and use these as our indicators and for control.
The choice of instrumentation requires:
knowledge of what properties change,
how they can be measured
the availability of suitable instrumentation.
There is no one right way to do anything, though perhaps, many wrong ways.
Instrument options What are our choices? We might consider these options because we know there are process instruments available for all of them. There may be other options available.
Choosing between them We may well decide that though the viscosity change is of a greater magnitude than the density change viscosity is not a good choice because most viscometers are only 1% full scale accurate, while we know we can get density meters that measure to +/0.15kg/m3. How do we choose? We need to know what accuracy of concentration measurement we need and then see what each instrument can yield based on:
the change in the property with concentration
the resolution of the instrument
Narrowing the field If we choose density we then have a range of different techniques to choose between.
force balance (someone, somewhere is still making the Gravitrol)
If we choose vibrating element we have a choice of types,
If we choose tube, do we use a dedicated density meter or a mass meter? Which manufacturer? In the end, choices may be limited by what is offered from trusted/approved suppliers.
There is not time in the day, nor days in the week, for any one engineer to evaluate all the options. Nor would it be possible to choose "the best" or the "most cost effective".
It is usually possible to find a solution that meets most, if not all the target requirements and within budget from an approved supplier. It will not necessarily be the best performing or the least expensive. A great many application solutions do work but are not the "best" way to do things, nor the cheapest nor the most accurate. That doesn't mean that anyone has made any bad choices, but good choices within the limits of the resources available to find a solution. The best choices are often made based on experience, our own or someone else's, and the trust we place in the chosen instrument companies to provide, as they increasingly do, a solution and not just an instrument.
Example In this example of caustic dilution, mass flow and density solutions are both commonly used as described.
We have two approaches: 1) Predictive: batch blend ratio determination 2) Feedback: process measurement.
The predictive approach depends on knowing the original qualities of the ingredients and on being able to predict the relative proportions of each that, when mixed or blended together, will yield a product that meets the target specification.
In an approach using mass flow meters, Raw concentrated caustic is stored in a bulk tank. The mass meters either batch the raw caustic and water into a mixing tank in the appropriate mass ratio or they control the mass flow rate ratio of the raw caustic and water into a pipeline with a static mixer.
In theory, the pipeline blending approach will allow on-demand blending. Simply open the demand valve and the mass meters will maintain the flow of water in the required ratio to the flow of caustic.
In practice, predictive methods depend on assumptions that often do not hold true. The strength of the raw caustic may be too variable and/or may stratify in storage. Even if a more tightly controlled quality of raw caustic is purchased, which is more expensive, the variations may still be unacceptable.
The alternative (or the extra) stage is to deliver the blended caustic into an intermediate tank where the quality is tested and corrected by the addition of more water or raw caustic. In mass meter schemes, a mass meter may be used as a density meter.
Note that many early mass meters, used as density meters, lacked the density performance for the feedback method described later. Mixing is an exothermic reaction. To use an early mass meter as a density meter in this application would be beyond its capabilities. Better was to pre-mix the approximate strength, allow the contents of the tank to cool, and then make the density measurement (in a pumped re-circulation loop).
The feedback method using dedicated density meters is also well proven and preferred by many. Modern mass meters have much improved density performance (based on improved calibration and software enhancements) than previously and can now be considered for use as dedicated density meters. Note that not all mass meters are equivalent in performance or potential performance.
The 43% raw caustic and the water are pipeline blended using a static mixer. Downstream of the static mixer is a density meter.
It measures the density at the stream temperature and the temperature. Using a referral method e.g. matrix ratio referral, it determines the density at the reference temperature (15[°]C). Dependent on the resolution required, the density at 15[°]C may be assumed linear with the concentration. If the system is only to produce one concentration, e.g. 23%, this is fine. If the system allows the operator to select the strength required then a further algorithm is required as concentration is not linear with base density.
The concentration measurement produced is now used to modulate the water flow rate against the raw caustic such as to maintain the target quality.
In this approach, using good accuracy density meters, there is no need for an interceptor tank, nor any post mixing correction.
This approach is ideal for on demand blending and for variable set-point.
The calculation steps are:
measure density of the mix and the temperature.
Using matrix-referral (i.e. stored curves of temperature Vs density for different concentrations) find the density at 15degC.
calculate the concentration from the density at 15degC. Assume linear if only one target set-point (and calibrate this point) or apply or configure the algorithm to correct for the non-linear relationship between density and concentration.
modulate the control or mixing valves accordingly
When choosing a method and instrumentation, be sure that the process conditions are considered, the nature of the reaction and any implications this may have, the calculation steps and the target specification.
Compare instruments based not on their density measurement performance, or mass flow rate accuracy, but on the guaranteed ability to meet a concentration accuracy specification. Be very carefull when comparing instruments based on their declared density measurement accuracy. Many instuments appear to perform the same at the calibration conditions but these are not the process conditions and it is here that differences emerge. This is particularly true in exothermic reactions where temperature plays a big part.
Beware of bubbles. Exothermic reactions may promote the release of dissolved gases, or even stem formation. A pressurized pipeline with inline mixing is often a more satisfactory approach. Most sensors are vulnerable to bubbles. Where bubbles are a significant problem, the choice of instruments may be more limited. (see the FAQ on Ethylene Glycol dilution).
Useful Resource Pioneer (now Olin Chlor) publish extremely useful data on NaOH blending including density and viscosity Vs concentration, an enthalpy diagram and a guide on calculating the temperature rise due to the exothermic reaction.
If only more companies provided such useful information!