5.2.10 Control Methods

Control Methods Introduction to Control Control Methods ensure that process improvements are maintained over time and that performance does not slip back to previous levels. In the DMAIC roadmap, Control is the phase where the improved process is stabilized, monitored, and handed over to routine management. Control Methods focus on: - Defining what must be controlled and why. - Selecting appropriate monitoring and response methods. - Designing process and product controls. - Documenting, training, and transferring ownership. - Reacting effectively when performance drifts. The objective is not just to “freeze” the process, but to create a disciplined system that sustains gains while allowing controlled, data-based adjustment. --- Control Plan Fundamentals Purpose of a Control Plan A Control Plan is a structured document that summarizes how critical aspects of a process or product will be monitored and controlled. It links the improved design of the process to day-to-day operation. Key purposes: - Maintain process performance at the desired level. - Detect shifts or trends before defects reach the customer. - Clarify responsibilities and reaction steps. - Standardize control across shifts, locations, and teams. Elements of a Control Plan Typical elements include: - Process step – Clear description of the activity or operation. - Characteristic – What is being controlled (e.g., dimension, cycle time, error rate). - CTQ / specification – Target, specification limits, or performance requirement. - Measurement method – How data is collected (instrument, checklist, system). - Sampling strategy – Frequency and sample size for checks. - Control method – Chart type, alarm limits, visual controls, mistake-proofing. - Reaction plan – Specific actions when out-of-control or out-of-spec is detected. - Responsibility – Who measures, who reviews, who reacts. - Documentation – Forms, systems, or logs where results are recorded. A Control Plan should be practical: detailed enough to guide consistent behavior, but simple enough for daily use. --- Selecting Key Characteristics for Control Critical To Quality Characteristics Not all variables can or should be controlled with equal rigor. Control Methods emphasize focusing on characteristics that have the greatest impact on customers and performance. Key types: - CTQ (Critical to Quality) – Directly related to customer requirements. - CTP (Critical to Process) – Key process variables that affect CTQs. - CTC (Critical to Cost) – Drivers of cost, waste, or rework. Selection criteria: - Strong statistical relationship to the outcome (from Analyze/Improve). - Historical instability or sensitivity to change. - High risk or high consequence if performance drifts. - Feasibility of measurement and timely response. The Control Plan centers on these critical characteristics while using lighter control for less impactful variables. --- Statistical Process Control (SPC) Basics Purpose of SPC in Control Statistical Process Control provides a data-based way to differentiate between inherent process variation and special causes that require action. It supports Control by: - Providing early warning signals of process change. - Quantifying process stability and capability over time. - Enabling structured reactions to variation. SPC does not eliminate variation by itself; it provides a framework for monitoring and decision-making. Common vs Special Cause Variation Understanding variation types is essential: - Common cause variation - Natural, random, inherent in the current process. - Present when all known conditions are normal. - Reduced only by changing the process design or system. - Special cause variation - Arises from specific, identifiable sources (e.g., equipment fault, wrong material). - Appears as unusual patterns or outliers on control charts. - Addressed by investigation and corrective action. Control charts are designed to detect special causes while avoiding overreaction to common cause variation. --- Control Charts: Concepts and Selection Control Chart Fundamentals Control charts plot data over time with: - A center line – Usually the process average. - Upper and lower control limits (UCL/LCL) – Statistical limits, typically ±3σ from the mean. - Individual data points or subgroup statistics. Key ideas: - Points within control limits with random pattern → process is stable (common causes only). - Points or patterns violating control rules → process is unstable (special causes present). Control limits are not specification limits. Specifications describe what is acceptable to customers; control limits describe how the process behaves. Variable vs Attribute Charts Control chart selection depends on data type and sampling: - Variable data (continuous) - Measured on a scale (time, length, weight). - Examples: - X̅-R chart – For subgroup means and ranges (moderate subgroup size, e.g., 2–10). - X̅-s chart – For subgroup means and standard deviations (larger subgroups). - Individuals-Moving Range (I-MR) chart – For single observations or low-volume processes. - Attribute data (discrete) - Counts or classifications (defects, defectives). - Examples: - p chart – Fraction defective, variable sample size. - np chart – Count of defectives, constant sample size. - c chart – Number of defects per inspection unit, constant area of opportunity. - u chart – Defects per unit, variable area of opportunity. Correct selection ensures valid interpretation and appropriate sensitivity to change. --- Interpreting Control Charts Basic Control Chart Rules Control Methods rely on standard rules to detect special causes. Common rules include: - Any point outside UCL or LCL. - A run of several points on one side of the center line. - A trend of points continuously increasing or decreasing. - Cyclical or repeating patterns. - Unusual clustering of points near a limit or near the center. When rules are violated: - Do not adjust the process blindly. - Investigate possible special causes. - Confirm cause-effect with data when practical. - Implement corrective and preventive actions. The reaction plan in the Control Plan should specify which rules trigger which actions. Limits vs Specifications Important distinctions: - Control limits - Calculated from process data. - Describe what the process is currently doing. - Used to detect special causes. - Specification limits - Defined by customer or design requirements. - Describe what the process should achieve. - Used to judge acceptability of output. A process can be: - Stable but incapable (in control yet frequently outside specs). - Unstable but sometimes meeting specs (out-of-control but often inside specs). Control Methods aim first at stability, then at maintaining or improving capability. --- Process Capability in the Control Phase Capability Indices: Cp, Cpk, Pp, Ppk Although capability is often emphasized in Improve, capability monitoring is part of Control. Key indices: - Cp, Cpk (short-term capability) – Based on within-subgroup variation. - Cp – Compares process spread to spec width; assumes centered process. - Cpk – Accounts for centering; lower of CPU and CPL. - Pp, Ppk (long-term performance) – Based on overall variation. - Pp – Overall spread relative to specs. - Ppk – Overall centering. Control usage: - Use capability indices as baseline and future checkpoints. - Track capability periodically to confirm improvement sustainability. - Investigate significant deterioration in Cp/Cpk or Pp/Ppk. Capability indices are meaningful only when the process is stable or at least reasonably controlled. --- Reaction Plans and Standardization Structured Reaction Plans A reaction plan defines what to do when control charts or other controls signal a problem. It transforms data into consistent action. Typical reaction plan contents: - Trigger condition - Out-of-control signal on a specific chart. - Measured characteristic outside specification. - Missed sampling or measurement. - Immediate actions - Contain suspect product or transactions. - Stop or isolate affected portion of the process where necessary. - Inform designated person or role. - Root-cause checks - Pre-defined quick checks (e.g., material lot, instrument status, setup verification). - Escalation path if cause is not immediately found. - Corrective actions - Process adjustment instructions. - Rework or disposition of affected output. - Documentation - Logging of occurrence, cause, and actions taken. - Follow-up on recurring patterns. Reaction plans are critical to converting Control Plans and SPC into effective, repeatable control. Process Standardization Standardization prevents drift and variation in how the process is run. Control Methods emphasize: - Standard work instructions - Clear steps, sequence, and key points. - Defined tools, settings, and materials. - Checklists and job aids - Setup and shutdown checklists. - Pre-flight checks for critical tasks. - Version control - Ensuring only current procedures and instructions are used. - Controlled updates when improvements are made. Maintaining alignment between Control Plan, standard work, and actual practice is a central task within Control. --- Mistake-Proofing (Poka-Yoke) in Control Role of Mistake-Proofing Mistake-proofing designs the process so that errors are prevented, detected immediately, or made impossible. It strengthens Control by reducing reliance on human memory and vigilance. Applications: - Prevention – Designing steps so the wrong action cannot occur (e.g., connectors that only fit one way). - Detection – Automatic detection of incorrect conditions with stops or alarms. - Guidance – Physical or visual guides that make the correct action natural and the incorrect action difficult. Mistake-proofing, combined with SPC and standardization, creates layered defense against process deterioration. Evaluating and Maintaining Mistake-Proofing For sustained control: - Assess effectiveness: Does the device truly prevent or detect the targeted error? - Ensure maintainability: Can it be easily checked, cleaned, or calibrated? - Include in Control Plan: - Description of device or feature. - How often it is checked. - What to do if it fails or is bypassed. Mistake-proof devices themselves require control to stay reliable. --- Visual Management and Process Control Visual Controls Visual controls make process status, performance, and standards immediately visible. Common forms: - Status indicators – Lights, boards, dashboards showing whether process is on target. - Visual standards – Pictures, samples, layouts showing correct conditions. - Limits and markers – Lines, labels, color coding for positions, quantities, or priorities. Within Control Methods, visual controls: - Support adherence to standard work. - Provide quick feedback to operators and supervisors. - Promote timely reaction when conditions deviate. Integrating Visual Management with SPC Visual management should not replace data-based control but complement it: - Display key control chart summaries or signals on visual boards. - Use visuals to show: - Current vs target performance. - Recent out-of-control events and actions taken. - Status of improvement actions and preventive maintenance. This integration helps keep Control Plans active and visible rather than hidden in documents. --- Measurement System Control Ongoing Measurement System Evaluation Control Methods require continued confidence in measurement systems used for control charts and decisions. Key practices: - Periodically review measurement system performance - Repeatability and reproducibility (based on prior MSA). - Stability over time using control charts on measurement checks. - Linearity and bias where relevant. - Use reference standards - Check gauges or instruments with known standards at defined intervals. - Record and react to drift or out-of-tolerance conditions. If the measurement system deteriorates, control charts and capability metrics become unreliable, undermining Control. Reaction Plan for Measurement Issues Include in Control Plan: - Triggers for recalibration or instrument replacement. - Procedures for quarantining data collected with suspect instruments. - Reassessment steps when significant measurement issues are discovered. Maintaining measurement integrity is essential to all other Control Methods. --- Documentation, Handover, and Training Control Documentation Control requires clear, accessible documentation that reflects the improved process. Essential documentation: - Control Plan and associated reaction plans. - Updated process maps or flow descriptions. - Standard work instructions and checklists. - Control charts setup: chart type, sampling, limits, rules. - Process capability summary and target performance. Documents must be aligned and consistent so that operators are not faced with conflicting guidance. Training and Handover Sustaining control depends on transitioning ownership to those who run and manage the process. Key elements: - Targeted training on: - What is being controlled and why. - How to take measurements and read charts. - How to execute the reaction plan. - Practice using actual control charts and tools. - Clarification of: - Daily responsibilities. - Escalation paths. - Expectations for documentation and communication. Handover is complete when process owners can maintain the Control Methods without external support. --- Sustaining and Improving Controls Monitoring and Review Control is not static; controls should be periodically reviewed. Routine reviews focus on: - Stability of key characteristics over time. - Trends in defects, rework, or capability indices. - Frequency and nature of out-of-control events. - Effectiveness of reaction plans and standard work. If controls are too loose, problems slip through; if they are too tight or burdensome, they may be bypassed. Review balances effectiveness with practicality. Continuous Improvement of Control Methods Even after Control is established: - Simplify measurements and charts where possible. - Strengthen mistake-proofing to reduce reliance on human detection. - Adjust sampling frequency as process stability improves. - Incorporate lessons learned from incidents into updated Control Plans. Control Methods support both sustaining gains and enabling incremental improvement over time. --- Summary Control Methods ensure that improved processes remain stable, capable, and aligned with customer requirements. Core components include: - Clear Control Plans that define what is controlled, how it is measured, and how to react. - Appropriate use of SPC and control charts to distinguish common from special causes. - Ongoing monitoring of process capability to confirm sustained performance. - Structured reaction plans and standardized work to respond consistently to variation. - Mistake-proofing, visual management, and robust measurement systems to prevent and quickly detect problems. - Thorough documentation, training, and handover so that process owners can manage controls daily. Together, these methods transform one-time improvements into long-term, reliable performance.