A note before we start
This article describes a batch process example that is entirely hypothetical. The facility, the product and the results are invented to illustrate the thinking involved. No real company or data is represented, and nothing here is engineering advice.
The setting
Picture a small facility that makes a liquid cleaning concentrate in batches. Each batch follows a recipe: fill a mixing vessel with water, add ingredients in a set order, heat to a target temperature, mix for a set time, cool, sample, adjust if needed, and transfer to storage. A few product variants use slightly different recipes.
Today the process is mostly manual. Operators open valves by hand, read gauges, watch the clock and write down values on paper sheets. The facility’s concerns are batch-to-batch variation, the effort of keeping records, and the risk that someone skips or reorders a step on a busy shift.
Why a batch process is a good automation candidate
Batch processes are a sequence of steps with conditions between them. That structure maps naturally onto automation. A recipe can be stored once as a series of phases, each of which waits for a condition (a level reached, a temperature held, a timer complete) before moving on. The same equipment can run several recipes by selecting different parameters.
The goal is rarely to remove people. It is to make each batch follow the approved recipe consistently while giving operators better information.
Decision 1: What to automate first
The team could automate everything at once, but they choose to begin with the parts that are repetitive and well understood:
- Filling to a target level
- Heating and holding at temperature
- Timed mixing
- Transfer to storage
Judgment-heavy tasks stay manual at first, for example adjusting the batch after a sample result. Over time some of that might be automated, but only once the team understands it well enough to write it down precisely.
Trade-off: a partial approach leaves some manual steps and some variation. It also reduces the risk of automating a step nobody can fully describe.
Decision 2: Instruments and final elements
Automating a recipe requires dependable measurements and controllable devices. Tank level measurement tells the system when to stop filling. Process control sensors measure temperature. Flow meters can confirm how much of an ingredient was added. On the output side, automated control valves and valve actuators replace hand-operated valves, and the mixer and pumps are started and stopped by the controller.
The team reviews what the plant already has. Some existing gauges are only local indicators and would need replacing with devices that send a signal. They also decide which valves need position feedback so the system can confirm a valve really opened, not just that it was told to.
Trade-off: extra instruments increase cost and maintenance, but without feedback the system can only assume things happened.
Decision 3: Recipe structure and interlocks
The controller organizes the process into phases such as “charge water,” “heat,” “hold,” and “transfer.” Recipes then become lists of phases with parameters (target temperature, hold time, ingredient quantity). Changing a product means editing parameters, not rewriting logic.
Interlocks are added to prevent unsafe or wasteful actions: for example, not starting the heater unless the vessel level is above a minimum, or not opening a transfer valve if the receiving tank is nearly full. Critical safety functions are treated separately from ordinary control and designed according to applicable standards and by qualified people.
Decision 4: Alarms and operator screens
Operator interface screens show the state of the batch, the current phase, key measurements and any alarms. The team is cautious here. Too many alarms cause people to ignore them, so alarms are limited to conditions that need action, with clear wording.
Operators also keep the ability to pause a batch, hold a phase and take manual control in a controlled way. The system supports their judgment; it does not replace it.
Decision 5: Batch records
Because the system logs events, the old paper sheet can be replaced by an automatic record of when each phase started and ended, what values were reached, and which alarms occurred. This saves time and reduces transcription errors. If the facility works in a regulated field, the record-keeping requirements would need careful review with qualified specialists.
Commissioning and training
Before running real product, the team tests the logic with the vessel filled with water only. They check each interlock, simulate sensor failures and confirm what the system does when power is lost or a signal drops out. Lockout tagout procedures apply whenever people work on the equipment. Operators are trained on the new screens and on what to do in manual mode.
What the example illustrates
- Batch steps and conditions suit automation well.
- Instrument quality limits automation quality.
- Interlocks and alarms deserve as much design effort as the main sequence.
- Records become a by-product of running the batch.
- People still decide when the process is behaving unexpectedly.
In a real facility the specifics, and the results, would differ and cannot be guaranteed.