Business Survival in the Short Term
To remain competitive in the short term, small-medium manufacturers (SMMs) are continuously (1) reducing costs; (2) improving quality; (3) increasing throughput; (4) enhancing safety … and that is just the short list! With such day-to-day pressures and the associated daily fire fighting it brings, it is no wonder that SMMs find it difficult to carve out time for longer range strategic planning. One aspect of a long-range plan that can get short-changed is future technology investments on the factory floor. We are in an era of accelerating technology convergence that has led to an enormous avalanche of decisions to make about technology adoption and vendor selection. This era has many names, including Industry 4.0 but the one that is most familiar is Digital Manufacturing.
So what is Digital Manufacturing and why should I care?
Digital Manufacturing codifies what has been called a “smart factory.” Within these smart factories, cyber-physical systems monitor physical processes, supporting a virtual copy of the physical world and enabling decentralized decisions. Digital Manufacturing is not a collection of technologies per se, but rather a set of design principles to guide the technology adoption for manufactures who need to “future proof” their investments to remain competitive in the longer term. I suggest that people think of it as a “North Star”. Absent of a strategic direction such as Digital Manufacturing SMMs are at risk of making each factory floor technology implementation as a one-off decision and soon find that they have sub-optimized their overall investment. The four principles of Digital Manufacturing are as follows:
- Interoperability: The ability of machines, devices, sensors, and people to connect and communicate with each other via the Internet of Things.
- Information Transparency: The ability of information systems to create a virtual copy of the physical world by enriching digital plant models with sensor data. This requires the aggregation of raw sensor data to higher-value context information.
- Technical assistance: The ability of assistance systems to support humans by aggregating and visualizing information comprehensible for making informed decisions. The ability of cyber-physical systems to physically support humans by conducting tasks that are unpleasant, exhausting, or unsafe for human co-workers.
- Decentralized decisions: The ability of cyber-physical systems to make decisions on their own and to perform their tasks as autonomously as possible. Only in the case of exceptions, interferences, or conflicting goals, tasks are delegated to a higher level.
How might my strategy need to change?
While lean, six-sigma and other shop floor practices are critical, employee training alone will not be sufficient to compete going forward. Advanced automation technology adoption by SMMs is occurring at an increasing rate. The emerging “Digital Manufacturing” era represents a convergence of core technologies including smart sensors, big data, cloud computing, collaborative robots, machine learning, advanced vision systems and others which have resulted in an acceleration of new factory automation solutions.
One strategy is to layer “Digital Manufacturing” principles together with prudent project selection guided by characteristics of:
- Low initial cost
- Easy to implement
- Fast ROI
- Low risk
However identifying these projects is challenging due to the uncertainty of assessing how difficult or costly a project is to implement, how much operational risk is involved or how fast will an ROI be achieved.
Technology Adoption Risk
Relative to management attention and focus, many SMMs are consumed with day-to-day commitments and operational “fire fighting.” Given this fact, SMMs are vulnerable to three decision “failure modes” which can critically and negatively impact their technology adoption:
- Decisions overlooked
- Decisions made poorly
- Decisions implemented poorly
“Shortlisting” your Key Decision
Connected Factory Global (CFG) is focused on understanding four aspects of Digital Manufacturing:
- What are the needs and pain points of the SMM?
- What are the best fit technologies which can solve those pain points?
- Who are the best solution providers who can implement these technologies?
- What are the key decisions the SMM needs to make for successful implementation?
We spend our time “boiling the ocean” so that our SMM customers do not need to. We boil the ocean by visiting factories, interviewing plant managers, line supervisors, and controls engineers, attending trade shows, talking with system integrators and solution providers and reviewing industry publications, case studies and white papers. Our value-add lies in the methods we use to curate this information into “Decision Patterns” which the SMM can leverage to jump start their Digital Manufacturing implementation. Our decision patterns help the SMM by identifying:
- Which Digital Manufacturing technologies represent the “low hanging fruit” (Low initial cost, easy to implement, fast ROI, low risk)?
- What is the shortlist of key decisions to consider for each technology?
- What is the shortlist of alternatives to consider for each decision?
- What is the shortlist of criteria to consider by which to choose the best alternative?
Why Automated Inspection?
When we talk about automated inspection, we are usually referring to in-line 100% inspection as opposed to off-line measurement room inspection. There are scenarios that may be considered as automated off-line such as a programmable CMM or a “next-to-line automated gauge but at CFG we are focusing on automated in-line 100% inspection stations and the challenges SMMs may encounter when justifying, specifying, installing and successfully operating and maintaining over the long term.
Whether dedicated inspection stations or process stations with inspection capabilities there are clear advantages to automated 100% inspection over off-line sampling. Labor savings and timeliness of detecting a process breakdown or product defect are three examples.
Another aspect of in-line automated inspection is that it is predominantly non-contact to avoid wear. Sensor technologies described as laser triangulation, capacitive, magnetic, confocal, eddy current, and 2D and 3D vision have all proven themselves in countless applications. These sensors all have specifications that describe precision and parameters such as linearity, resolution, measurement range, sampling frequency, and temperature stability.
When specifying an automated measurement station, there is a hierarchy of automation elements and related decisions. At the bottom are the sensors, actuators, transport and fixturing. Often a mixture of sensor technologies is applied to execute the desired inspection tasks. In the middle are the controls, interfaces, and software to connect the sensors and analyze, visualize or archive the inspection results. And at the top are the strategies for handling rejected parts, maintaining the system and managing failure modes.
For an SMM, depending on experience, resource availability, and strategic goals, automated inspection systems may be purchased turn key, often as part of a new process line, or may be implemented with internal resources. In either case, applying experienced-based decision patterns that include exemplary criteria and alternatives as starting points can assure the key decisions are not overlooked and can increase confidence that decisions are made well and implemented following a “First Time Right” philosophy.
Decisions to not overlook
As an example, with automated inspection some critical system life cycle level decisions include:
- System reliability monitoring methods
How do we know for certain at any time that the inspection system is operating correctly? Not passing bad parts? How to achieve and keep 100% confidence in the system?
- System changes control procedures
How do we know if someone made a change to the system that might affect its performance or capability? How do we confirm that a change does not adversely affect the system performance or capability? Can we easily reverse a change? Can we regression test program changes on archived data sets to assure the changes do not adversely affect the system performance or capability?
- System failure mode mitigation
What happens if the in-line inspection system goes down? Do we stop production? Do we bypass the inspection? Does the system identify a malfunction to alert operators? How to troubleshoot the problem? What is the offline backup plan?
- Training and retaining key skills
When we outsource the inspection system design, integration, and installation how much inside expertise do I need to effectively operate and maintain the system? Who needs to be trained to work with the system? How do we maintain the expertise over time and personnel changes?
- Data management strategies
How much data does the system produce? Who needs the data and in what form? What part/form of the data do we need to archive? Where do we archive the data? For how long? How do we access and use the archived data?
- Defective Material Handling Method
How do we handle rejected parts? How do they get marked, tracked, removed? Rework and then reintroduce into the line to re-inspect? Rework and inspect offline?
Decision Example – System Reliability Monitoring Method
When an inspection system is commissioned, typically a carefully planned acceptance test is performed including, in the case of gauging, a gauge repeatability and reproducibility (GR&R) evaluation to verify the system’s performance. A selection of parts that exhibit the critical dimensional range or failure modes that the system has been designed to inspect is used for the acceptance test evaluation. This can be a very time and labor intensive effort and may involve multiple trials and optimizations of the system.
For gaging dimensional measurement applications both repeatability and accuracy are important. Repeatability refers to how accurately a measurement can be duplicated. Accuracy is the measure of how well the instruments readings match the actual dimensions in reference to a calibration standard. Usually, an inspection systems accuracy and repeatability combination should be 10 times smaller than the tolerance range associated with the dimensions being checked. This is to assure the inspection system error does not influence the results.
After the system passes the acceptance test and is put into production there remains the question of how do we know if the system is maintaining its performance?
There are many opportunities for a system’s performance to become compromised. Some of the failure modes that need to be considered include:
- an inadvertent software parameter change by an inexperienced user
- a change in behavior following a software update
- a mechanical shift or drift of a sensor
- dirt on the sensor or lighting
- a mechanical wear or breakdown of a fixture or pallet
Therefore a method is needed to periodically verify the performance of the system to assure that good parts are not rejected as bad and defective parts are not passed as good.
Overlooking this decision, or choosing the wrong method or poorly implementing the method can result in high costs of scrapped parts, costly labor and customer satisfaction quality issues. Eventual loss of confidence in the system can lead to inefficient and expensive production workarounds. Implementing a new method after the system is installed can be difficult and expensive.
Figure 1 – Decision Breakdown Structure
In this case when we consider the four Digital Manufacturing principles we think of:
Interoperability: the controls, sensors, part tracking, etc. should be integrated to interface to the plant information system.
Information Transparency: the data from the system, if compromised or suspected to be compromised, needs to be clearly tagged as such in order to filter at a higher level.
Technical assistance: automation of the method is desired to minimize the manpower and assure the reliability.
Decentralized decisions: the goal of a completely autonomous system may be very challenging however the system should manage itself such that when a performance compromise is detected steps for resolving the issue are built into the programming.
These principles inform the criteria that are used to assess the available alternatives.
How do we maintain confidence that the Automated Inspection system is functioning properly?
In choosing a method, the first step is to list and rank the criteria that apply. The most critical criteria for the reliability monitoring/verifying method is that it is robust, reliable and will not fail in it’s intended purpose. Very important criteria include implementation costs and the Digital Manufacturing principles derived criteria of automated, decentralized, and transparent.
Key evaluation criteria for this decision:
- High reliability
- Potential for undetected failure of method
- Implementation cost
- Simple to manage (automated)
- Decentralized decisions
- Information Transparency
- Inspection system availability
- Operating skills
- Operating cost
- Operator safety
Next, the list of alternatives to consider is identified from evaluating the manufacturing process design, past experience and investigating similar systems. The alternatives identified for the reliability monitoring/verifying method include:
- Alternatives to consider
- Daily master part test
- Monthly Gauge R&R test
- Correlate to off-line inspection gauge
- Statistical Analysis
- In station fiducials
Within the CFG decision pattern database tool, the decision details are captured and ranked. The power of the decision database is the curation of the applicable criteria and alternatives and their associated metrics and risks over many projects. This accelerates the decision process, increases confidence and minimizes the risk of a poor decision.
Each alternative is evaluated against each criterion per objective and threshold metrics which provides a weighted score. This process assures deeply considered criteria and metrics. The alternatives of In-line fiducials and Statistical Analysis while scoring very high in the Digital Manufacturing principles are rejected due to falling below the threshold for cost and reliability respectively.
It Can be helpful for analysis and group discussion to visualize the results via a spider chart where the longer spokes are the more important criteria and the distance along the spoke is how well the alternative meets the criteria.
Of the remaining alternatives testing with master parts and testing via off-line equipment score highest, however, each has implementation risks that need to be considered.
Again, curated alternatives with considered and refined risks and risk mitigation tasks inform the decision. The most impactful risk with using a Master part test is the chance that it does not capture all possible failure modes. Mitigation of this risk requires careful planning of the Master Parts to be used in the test. Finally, the top alternatives can be compared on a Tornado chart visualization of scores for each criterion and the risks associated with each alternative can be compared.
For this example, both the Master Part Test method and Off-line test compare favorably, however, the risks associated with the Master Part Test are considered easier to mitigate leading to a decision.
About Connected Factory Global
Connected Factory Global (CFG) is a professional research, data intelligence, and advisory services firm focused on helping small and medium-sized manufacturers (SMMs) be successful in the new era of Industry 4.0 (e.g., “smart factories” and “connected supply chains”).