resilient systems design

Design Resilient Systems

Core Competency Resilient Systems

How do Complex Systems Fail?

There is a gap between the system as imagined and the system as is.

Imagined

System

System

as IsDrift

Imagined Systems are:

1. Static

2. Deterministic

3. Result of design

development

4. Assumes system

discipline.

As Built Systems are:

1. Stochastic: Driven by random

variables

2. Constantly require maintenance

3. Hard to predict.

Source: Richard Cook


How do Complex Systems Fail?

Why does drift occur?

Imagined

System

Random Events

Weakness in

System Design

System Change

Normal System

Variation

Increased costs

Time pressure

Safety problems

Quality problems


How to improve complex systemsRichard Cook

Resilient System Design Controls

Imagined

System

Random Events

Weakness in

System Design

System Change

Normal System

Variation

1. Create process for

continuous system

maintenance.

2. Reveal system controls

to operators

3. Reveal the leverage in

the system –lift points

4. Simulate failures

5. Create prevention

methods

Source: Richard Cook

Root Causes


Where do we apply Resilient Systems Design?

Production Systems: Standards, Takt times, Setups, Maintenance, Training, Documents, etc.

Quality Systems: Process Controls, Measurement Systems, Corrective Actions, Training, Inspection Systems

Safety Systems: Machines, Building Systems, Support Systems, Training

Information Systems: Standards, Product Codes, Schedules, Routing, Customer information


Time for Planning, Quality

Safety, and Information

Time for production

For maximum profit we want to minimize the non-value added

activities --- but they must be executed in a robust way.

When the non-value activities have a failure they cause

rebound problems on production


System is in balance, Time available for production, information, planning, quality and safety

The four support systems have time to prevent production problems and are not causing production

Problems.

Market

force

Market

force

Market forces increase pressure: Time for production squeezes out time for information, Planning

Quality and Safety------ Risks of failures increase


When a system problem occurs it will blowback and erode production time

Increasing costs and affecting customers.

The focus is to create systems that can withstand the pressure and

Still minimize risks.


Random Events will develop and push back against any part of the system.

The system must be able to detect these events quickly and deal with the

Problems before they get out of hand.


System Design Characteristics System Tools to accomplish Design

and to control processes

Create process for continuous

maintenance of the system.

Regular system audits.

Production system, quality system, safety

system and information system

Reveal System Controls to operators Train all employees

Audits, Corrective Actions, Process

Controls, FMEA’s, Variation reduction

Leverage the system

Identify the lift points

Ask the question– where is the leverage in

out systems. Focus on ongoing variation

reduction

Simulate failures Regular tests of the system.

Design and implement failure tests.

Fire drills, first aid response drills, gage

failure drills, machine failure drills

Develop a prevention system Apply QS APQP

New product process review

Control plans, FMEAs, MSA, Process

Controls, SPC, Pre-Control, Variation

Reduction


Four causes of drift in systems.

1. Weak systems design

2. Changes to the system

3. Normal variation in systems

4. Random variables.

What are some examples of actual activities that we can us to

address these problems.


Weak System DesignIn most factories some systems will be stronger than others.

Production is usually the strongest with safety the weakest.

Production System: Audit the system to TS standards, are there control plans, are

they being followed. Work instructions, training records, PM records, scheduling

systems

Quality System: Use the ISO, TS or AS audits. Are they being done regularly.

Procedures, work instructions, training records.

Safety System: Is there a safety system. Are OSHA audits done regularly. Are

there written procedures, work instructions, training records, are there

ongoing open safety problems.

Information system: Audit critical routers, are they correct? Is the coding correct.


System ChangeIn a strong system there will be a process to introduce new materials, new

parts, new machines, new processes.

Production System: There should be something that aligns with a QS APQP system for introducing new products. Is there a process and is it being followed.

Quality System: Are all the documents required to support new products in place? New measurement systems, new process controls, etc.

Safety Systems: Are new machines and processes reviewed for safety and hazardous material issues? Is this documented.

Information Systems: Is the information system adjusted for the new information? Is this change part of the APQP process?? Is it documented and reviewed. Is it tested?


Normal VariationVariation occurs in every process. In some cases we have data to

document it and in some cases we not. Our goal is to reduce the variation in our processes all the time.

Production System: Do we know the amount of variation in our machines, our standards, our cycle times, our setup times. We need to be improving these all the time.

Quality Systems: Do we have the skills to reduce our variation? Do we have the data from all of our processes?

Safety Systems: Need to gather data from our safety systems to begin to see variation.

Information Systems: Need to gather data from our information systems to begin to see the variation in our system.


Random Events that affect our systemRandom events can cause significant disruption. How we deal with them

will make our system stronger.

There are two key elements that help address random events in the system.

1. Ongoing monitoring of the system that reveals changes in the system and

the active measurement of system drift. This can be through SPC, audit

results or other active methods.

2. The active participation of trained operators who understand the monitoring

that they are doing. People need to be trained to react to random events

and to notify other people to get help to deal with them immediately.


Measuring Drift Managers should want to know how much their system as drifted from the

original design.

Monitoring this drift reveals a lot about the risk inherent in the business at any

point in time.

There are four key elements supporting production.

Quality, Safety, Planning and Information --- failure in any one of these

will rebound back on production.

Resilient systems theory tells us that any system is subject to the four key

problems. Random events, normal variation, poor system design or changes

in the system.


Measuring Drift

The economics of monitoring

Monitoring the value added portion of the factory is most important to generating cash flow.

The production system and the factory output get the most attention.

The primary metric is the standard costs income statement supplemented by daily KPI and cash flow.

The non-value added portions need just enough resources to insure that production is not affected by system failures.

There are institutional requirements such as ISO, TS and AS systems that require auditing to meet standard requirements.

There are OSHA requirements that require regular activities and should lead to regular audits

.


Four separate audits are conducted

on an ongoing basis and then presented at random

to management for review.


Alternative method to reduce costs of audits and to improve direct reporting of

drift.

Audit the plant by key product lines starting with the largest and progressing to

the smallest.

( Review existing ISO and Safety audits– separate work cell specific

material from plant wide material)

Create one audit that integrates Quality, Safety, Information and Planning into

one audit.


Resilient Value Stream Audit

Production: Takt time to output, work

standards, setup time, PM, 5S, training

records

Quality: SPC, Gage R&R, Cpk of

output, training records, Process

controls, corrective actions on key parts.

Information: Router accuracy,

standards, product coding review,

Raw material review, prior approvals.

Planning: Cycle times, on time delivery

Safety: Lockout, machine guarding,

ergonomics, noise, PPE, electrical,

training


How do Systems Fail?

Resources:

How Complex Systems Fail: Richard Cook Univ. of Chicago

Resilience Engineering: Erik Hollnagel, David Woods, Nancy

Leveson

A New Accident Model for Engineering Safer Systems

Nancy Leveson, MIT

resilient systems design

Leadership & Management