# Continuous, batch, discrete: the three worlds of production

> Every plant belongs to one of three families, based on what it makes and how. This distinction drives the choice of automation, tools and standards.

- Source : https://industryhub.site/en/learn/foundations/procede-vs-discret
- Track : foundations
- Level : Discovery
- Reading time : 5 min
- Language : English

## In brief

There are three types of production: continuous process (a flow that never stops), batch (by lots and recipes), and discrete (parts counted one by one). Each calls for different automation and different tools.

## Why this distinction comes first

Before talking about controllers or standards, you must know **what the plant produces and how**. Three families exist, and everything follows: the type of control system, the tools, the skills, the applicable standards.

| | Continuous | Batch (by lots) | Discrete |
|---|---|---|---|
| **Flow** | uninterrupted | by recipes, lot after lot | counted parts |
| **Examples** | refining, cement, paper | pharma, fine chemicals | automotive, electronics |
| **Typical tool** | DCS | DCS/PLC + recipes | PLC, robotics |
| **Dedicated standard** | — | ISA-88 | — |
| **Key metric** | stability, yield | batch traceability | OEE |

## The continuous process

A **flow that never stops**: refining, petrochemicals, power generation, paper, cement. Material enters, is transformed and leaves continuously. You regulate quantities (temperature, pressure, flow) at all times, with **control loops** by the hundreds. Typical tool: the DCS. Key challenge: stability — a stoppage propagates through the whole chain and is very costly. Start-up and shutdown are long, delicate, sequenced.

## The batch process

You make **by recipes**, lot after lot: fine chemicals, pharma, food, paints. You charge, transform per a recipe, discharge, clean, start again. Each lot is traced. Typical tool: DCS or PLC + recipe management. Dedicated standard: [ISA-88](/en/standards/isa-88), which structures recipes and phases. Key challenge: batch traceability and reproducibility.

## Discrete manufacturing

You produce **objects counted one by one**: automotive, electronics, machines, packaging. The part moves along a line, undergoes operations, is assembled, tested. Typical tool: the PLC and robotics. The key metric is **Overall Equipment Effectiveness (OEE)**, the product of three factors:

$$\mathrm{OEE} = \text{Availability} \times \text{Performance} \times \text{Quality}$$

A line available 90% of the time, running at 95% of its rate, with 99% good parts, shows an OEE of $0.90 \times 0.95 \times 0.99 \approx 85\%$. This single indicator reveals where productivity is lost: downtime, micro-stops or scrap.

## Many plants are hybrid

Reality often mixes all three: a food plant runs **batch** (cooking) then **discrete** (packaging). A refinery is **continuous** but its tank farm follows a management logic. Recognising which logic applies to which part of the plant is a core skill of the automation engineer.

The link between production and enterprise management (orders, stock, planning) is structured by the [ISA-95](/en/standards/isa-95) standard.