Essential Quality Tools for Process Improvement and Analysis

Posted on Oct 6, 2025 in Mathematics

Quality Tools for Process Control

Quality tools are essential instruments used to control a process, identify faults, improve risk analysis systems, and drive continuous improvement.

Brainstorming

Brainstorming is a technique where a group encourages every member to participate and generate ideas, focusing on quantity over complexity. It is crucial that the group avoids criticizing ideas during the generation phase, as only some ideas will ultimately be valid.

Requirements for Effective Brainstorming

  1. The groups should be small (3–8 people).
  2. Each member must know and fully understand the problem being posed.
  3. All ideas must be accepted without criticism.
  4. Participants can provide ideas that build upon one another and those already expressed.
  5. The duration of the meeting must be preset.

Phases of Brainstorming

  1. Clearly define the problem.
  2. Ideas are presented by participants.
  3. The group reflects on the ideas and selects the best solutions.

Cause and Effect Diagram (Ishikawa Diagram)

The Ishikawa Diagram, also known as the Fishbone Diagram, primarily helps locate the root causes of a specific effect. Causes are typically grouped into concrete categories (blocks). This analysis tool is very simple and versatile.

Steps for Creating an Ishikawa Diagram

  1. Select the effect that needs to be controlled or analyzed. This effect forms the trunk of the diagram, from which the branches representing the causes acting on that effect will emerge.
  2. For each main branch (category), group the resulting impacts or sub-causes deemed relevant.
  3. The causes must be sorted according to their importance relative to the problem being analyzed.

Pareto Chart (The 80/20 Rule)

The Pareto Principle, frequently observed in industrial processes and natural phenomena, states that the distribution of effects and their possible causes is not linear. Instead, approximately 20% of the sources cause 80% of the effects. The Pareto Chart visually represents this distribution, helping prioritize efforts on the vital few causes.

Histogram

A Histogram is used to visualize how a series of data is organized, helping determine the distribution and behavior of a variable associated with a process. It uses bars to represent the frequency distribution of a variable, grouped into determined intervals or bins.

Uses of Histograms

  • To see if the process follows the required specifications.
  • To observe if the data is dispersed around the desired value.

Sector Charts (Pie Charts)

Sector charts (or Pie Charts) are mainly used to represent percentages. Their form is circular with radial divisions. The angle (in degrees) corresponding to each segment is calculated using a rule of three, based on the fact that 100% corresponds to 360 degrees.

Control Charts

Control charts allow users to check if a process is stable over time relative to a specific variable that needs to be controlled. They typically mark Upper Control Limits (UCL) and Lower Control Limits (LCL). If the variable remains within these limits, the process is considered controlled; if it exceeds them, the process is out of control.

Types of Control Charts

  1. Control Charts for Attributes: Used to control a characteristic of the process, often binary (e.g., go/no-go, defective/non-defective).
  2. Control Charts for Variables: Used to control a measurable magnitude, variation, or dispersion.

Correlation Diagram (Scatter Plot)

A Correlation Diagram (or Scatter Plot) is used to detect the type of relationship (correlation) between two variables of a process. The starting point is the data pairs of the two variables whose relationship is being sought. One variable (X) is chosen for the horizontal axis, and the other (Y) for the vertical axis. The scale of each axis is set so that the limits coincide with the maximum and minimum values represented. Once the axes are completed, pairs of related values are plotted.

Types of Correlations

  • Linear Increasing (Positive Correlation): Increases in variable A produce increases in variable B.
  • Linear Decreasing (Negative Correlation): Increases in variable A produce decreases in variable B.
  • Linear Horizontal (No Correlation): Variation of A does not produce variations in variable B; B is independent of A.
  • Non-Linear Correlation: Variations in A produce different variations in B depending on the point in the data set.
  • No Correlation: In this case, it is not possible to adjust a line to follow the trend of the points; A and B are not correlated.