The Steps to a Successful POC
Author – Tim Smith
Your company is engaged in a proof of concept project where a single system is connected to the network and enrolled into a manufacturing operations management system. The steps to deployment and commissioning of the machine along with a framework of objectives is critical to a successful POC
1.Network Topology– An important aspect of any large scale information system is the layer 0, layer 1 and layer 2 topology. Concerns about performance, security, error fault detection and the effective utilization of resources mandate the assessment and creation of an effective topology. Include IT into the POC as a resource
a. Investigate networking options/scenarios- deployment, management, maintenance, security, scalability
b. Define conceptual layouts
c. Validate and select a layout (Much of this may already exist . Check with your IT group)
2. Machine selection– The motivation for selection of a machine asset(s) for a POC should be based on very clear outcome objectives. Objectives such as capturing operational events automatically, capturing specific process data, candidate for operator interaction based on operations. Possibly, auto-creation of jobs, error reason code identification and operator response lookup may be part of the assessment. Capturing machine specific capabilities such as mutli-spindle activity, pallets, tombstones, multi-part count, (don’t forget multi-operations producing multi-part counts), high speed part count, etc. Another objective approach is what position in a value steam the machine is. Is it a finishing machine which sets the throughput for a group of operations? Is it an identified constraint machine and collecting data to assist in constraint resolution to build/adopt a CI methodology.
a. Select criteria to determine objectives
b. Apply to and select machine candidates
c. Assess which candidates provide the widest range of information and connectivity by which topology objectives can be met
d. Select a machine asset(s) (Authors note: most successful POCs select a number of machine assets based on operation type. The consideration is not whether the ol’ dog in the corner will connect to a modern Machine Operations Management system, which it will, but whether the machine(s) selected also represent the bulk of machine assets) This point is also mitigated by the use of popular or open standards connectivity.
3. Information System Selection– Manufacturing information systems come in an ever growing range of offerings. From legacy SCADA and process mapping systems, MES and batch run systems, bolt-ons to mainstream systems such as ERP, emerging manufacturing operations systems and monitoring systems make up an ever growing list. Stakeholders will need to determine which function/features are the most desirable and prioritize them. The greatest challenge to the assessment and selection of a candidate for the POC is identifying the deliverables and quantifying the unexpected. With the rapid entry of products into the market all making the same claims it can be difficult to choose a winner. Things to consider:
a. System must be a service. No exceptions, executables in server side environments is a recipe for disaster.
b. System must be distributable by services across the compute environment.
c. System must be scalable. Is the solution an all-in-one application? This determines how costly scalability will be. If the architecture is distributable natively then it is scalable.
d. System must be expandable in feature and function. Can the platform support additional features? Is it modular? What is the ability to add views, screens, reports, web pages, metrics, custom data events, etc.? Not just user adjustments to a fixed view, but the ability to confugre brand new views with a user toolset.
e. System must be productized to reduce outside customization and development costs. You are buying a product not a consulting engagement. Many of the legacy systems such as Wonderware, Thing Worx, etc. are vastly customizable systems which is a made-to-order system requiring constant outside support.
f. Manufacturing, Information technology, system integration, mechanical and electrical engineering should all be IP silos from which the manufacturing operations system is built upon. A software only experience base usually caters to the visual aspect of the system at the cost of features and functions relevant to manufacturing and operations.
g. Vendor stability and longevity. How long has the vendor been around? Is it privately or publicly held? Does it have a reasonable personnel level to service you? Is it a thought leader in the industry? Does it have depth? Does it have an install base? Does it have success stories and use/business cases?
4. POC Execution– Once the topology is set, the machine(s) assets are selected and the information system selected it’s time to plan and execute a roadmap
a. List the common objectives for the roadmap. The objectives must be specific in nature and not general. They must be qualifiable. They must be quantifiable. The outcome from the POC must be successful in that stakeholders can apply it to the remainder of the plant.
b. Select internal champions to be responsible for allocating production, engineering and IT resources as necessary to execute the implementation
c. Arrange a kickoff meeting with the vendor to assess their action plan and distribution of responsibilities. The recipe for success is to make sure that your resources are committed and have departmental backing to get resources when needed during the POC.
d. Set a POC time limit. I would recommend 60 days. 30 days is too short to produce a good baseline. Longer than 60 days and the project looses steam and shiny things steal the attention of the stakeholders. (Note: within the time period selected make sure that there is a frequent reporting structure, possibly weekly to drive the POC to completion.)
Do it now. There is never a perfect time. Too busy and you won’t give the project enough attention. Too slow and the concern is that “we have better things to do with our time” syndrome sets in.
If you follow these steps you will have a successful POC and ultimately a Data Driven Manufacturing plant.