4. Quality Control

Definition of food quality control

The core mechanism of quality control is to maintain desired quality by:

1. Measuring of product and process parameters

2. Comparing outcomes with standard requirements

3. Deciding if adjustment is needed

4. Taking corrective action(s) to ensure final quality will meet requirements.

Quality control steps

Four basis quality control activities.

Measuring as a basis for quality control activites

Product measuring:

• Taking representative sample and measure product parameter

Process measuring:

• Checking the actual parameters

Types of measuring devices (Five total)

1. Off-line: Manual sampling, followed by transport (lab/test room) for actual measurment or analysis

2. At-line: Manual sampling followed by analysis or measurement at location

3. On-line: Automatic sampling followed by automatic analysis

4. In-line: Signal is observed by a sensor (thus no sampling) in the product stream and translated to an external signal.

5. Non-invasive: Measuring a signal in a product stream without physical contact with the product (portable ultrasonic flow meter)

Testing as a basis for quality control activities

• Testing: comparing outcome of the measurement with the target (norm) and tolerances (limits)

• The outcome of measurement can be e.g. a number (e.g. CFU/ml of pathogen) or a visual result (e.g. color)

• Testing can be done by people (e.g. for sensory test results) or be automated (e.g. weight measurements).

Steps in corrective actions

1. Regulating

2. Actual handling

3. Control circles

Corrective action: regulating

Regulating is the process of determining which corrective action is required: both the extent (large or small) and direction (positive or negative)

Corrective action: actual handling

• Corrective action → actual handling to bring system back in control

• Can be instrumental (e.g. temperature increase) or human handling (e.g. removal of the wrong products)

• Depends on accuracy of handling and documentation of actual correction

Corrective action: control circles

• Feedback:

  • The corrective action is taken after the trouble has occurred

  • This action adjusts the undesired situation based on measurements later in the process

• Feedforward:

  • The corrective action is taken before the trouble occurs

  • This action anticipates the expected situation based on measurements earlier in the process.

• Concurrent: combination fo feedforward & feedback control circles.

Why is food quality control being done?

• Because of variation

• Food products are ‘living systems’ which cause variation

• Quality is inversely related to variation; less variation → less scrap and reword, fewer line stoppages, happier customers

• Variation is a natural phenomenon → but performance of reliable process can be predicted within limits

• To regulate or improve a process we must understand causes of variation

structural vs. incidental variation

What kind of variation is this?

Structural variation

What kind of variation is this?

• Incidental variation

• Odd one out, not structurally happening.

Three types of sampling methods

1. Spot-check procedure

2. 100% inspection

3. Acceptance sampling

Acceptance sampling method

• Acceptance sampling is based on statistical principles

• Provides judgement of the risk of a decision

  • Producer’s risk probability that a batch is wrongly rejected

  • Consumer’s risk probability that the batch is wrongly accepted.

The principle is that you are not measuring the quality of the full batch, but you are to sentence the batch to either reject or accept it.

Steps in the acceptance sampling method steps

1. Sample n samples out of N (N = batch)

2. Determine the number of (non-) conforming items

3. Compare results with acceptance criteria

4. Decide on acceptance or rejection

Steps in the acceptance sampling method (Step 1)

Sampling plan aspects

• Location (e.g. incoming materials or final inspection)

• Which sampling plan (number of items from a batch)

• Which items to take (e.g. on the top, randomly)

• The type of data (e.g. the non-conforming or conforming, variable or attribute)

Steps in the acceptance sampling method (Step 2)

Determination of (non-) confomring items

• What type of analysis:

  • Visual inspection

  • Sensory evaluation (e.g. color, taste, flavor, odor)

  • Physical evaluation

  • Microbial analysis

  • Compositional analysis (e.g. protein and flavor concentration)

  • Enzyme analysis (e.g. proteinases, lipases, pectic enzymes)

Steps in the acceptance sampling method (Step 3)

Compare with acceptance criteria

Steps in the acceptance sampling method (Step 4)

Sampling plans (step 1)

How to design the sampling plan?

• Set the desired quality level: AQL/LTPD

• Usual errors chosen for sampling plan:

  • 5% chance of rejecting a good lot (AQL)

  • 10% chance of accepting a bad lot (LTPD)

• Caculate the sampling plan:

  • How many units ‘n’ to sample

  • Maximum number of defective items ‘c’

How do you visualize effect of sampling plan?

• Operating characteristic curve (OC)

  • Shows probability of accepting lots of different quality levels

  • Assists management to discriminate criteria of good and bad lots

  • Exact shape and location of the curve is defined by the sample size (n) and acceptance level (c)

Double sampling plans

Multiple sampling plans

Sequential sampling