Feb 3
Failure Modes and Effects Analysis (FMEA)
- Definition: Structured method for identifying possible failures in a product or process to assess their effects and find improvement opportunities.
- Failure Modes: Asks "What can go wrong?"
- Effects Analysis: Asks "What would be the consequences of failure for the customer?"
Example: FMEA for an ATM
- Function: Dispense cash
- Failure Modes:
- Out of cash
- Effects of Failure: Decreases customer satisfaction
- Causes: Low cash alert not functioning
- Recommended Action: Increase bill holder size
- Responsibility: Designer
- Jam issues
- Effects of Failure: Bills stuck, decrease in profitability
- Causes: Incorrect cash grouping
- Recommended Action: Riffle cash
Steps in the FMEA Process
- Identify Functions: Define what the product/process does.
- Identify Failure Modes: Determine ways failures could occur for each function.
- Identify Effects of Failure: Assess the consequences of failure from the customer’s perspective.
- Determine Severity Rating (S): Scale from 1 (insignificant) to 10 (catastrophic).
- Identify Causes of Failure: Find reasons behind the failure modes.
- Determine Occurrence Rating (O): Scale from 1 (extremely unlikely) to 10 (inevitable).
- Identify Control Mechanisms: Existing processes to detect failures.
- Determine Detection Rating (D): Likelihood of detecting the failure before customer impact.
- Calculate Risk Priority Number (RPN): RPN = Severity (S) × Occurrence (O) × Detection (D).
- List RPNs: In descending order to prioritize issues.
- Identify Recommended Actions: Steps to reduce failure severity or improve detection.
- Update Ratings: As actions are implemented, update S, O, D, and calculate new RPNs.
Demand Forecasting
- Definition: Estimate of expected demand for a specified time in the future, crucial for operations management.
Uses of Demand Forecasts
- Help design the system: Long-term planning.
- Plan medium-term system usage.
- Schedule short-term system usage.
Examples of Forecasts in Various Fields
- Accounting: Estimates for new products, profits, cash management.
- Finance: Equipment needs, funding requirements.
- Human Resources: Hiring and layoff planning.
- Marketing: Pricing, promotion, and strategy planning.
- Operations: Scheduling, capacity, inventory planning.
Common Features of Forecasts
- Based on the assumption that past causal systems will continue into the future.
- Rarely perfect due to randomness.
- More accurate when predicting groups rather than individuals.
- Forecast accuracy diminishes as the time horizon extends.
Elements of a Good Forecast
- Timeliness: Must be provided promptly.
- Accuracy: Must quantify the degree of accuracy.
- Reliability of method/software used.
- Expressed in understandable units.
- Must be documented in writing.
- Simple to comprehend and use.
- Cost-effective.
Steps in the Forecasting Process
- Determine the forecasting purpose.
- Establish a forecasting horizon.
- Gather and analyze historical data.
- Choose a forecasting technique.
- Prepare the forecast.
- Monitor the forecast.
Approaches to Forecasting
- Judgmental Approaches: Based on subjective inputs, considering human factors.
- Quantitative Approaches: Utilize numerical data to predict future values.
Judgmental Methods
- Executive Opinions: Combining insights from high-level executives.
- Sales Force Opinions: Insights based on customer interaction.
- Consumer Surveys: Gathering feedback directly from consumers.
- Historical Analogies: Using past demand data from similar products.
- Expert Opinions: Techniques like the Delphi method for consensus.
Time Series Models
- Sequence of observations collected at regular intervals.
Patterns in Time Series:
- Level: Stability, showing no significant trend.
- Trend: Directional movement over time.
- Seasonality: Fluctuations related to time periods.
- Cycles: Variations over longer durations due to economic factors.
- Irregular Variations: Anomalies not reflecting typical behavior.
- Random Variations: Noise after accounting for other patterns.
Time Series Techniques
- Naive Methods: Simple, low-cost but low accuracy.
- Averaging Methods: Include moving averages and exponential smoothing.
- Trend Models: Account for linear or non-linear trends.
Naive Method Explained
- Stable Series: Next point forecast equals the last data point.
- Seasonal Variations: Forecast equal to the last seasonal value.
- Trend Data: Forecast derived from the last value adjusted by trend difference.