What to Expect with Onsite and Remote Services
Most plants do not have a spare controls engineer sitting around waiting for a monitoring project. The work still has to happen: servers stood up, screens built, historian tags mapped, drawings updated, fiber pulled between the switchgear building and the control room. The question is who does it, where they do it from, and how much of your team's time it consumes.
Cutsforth's Automation and Control Services team supports power plants and industrial facilities across the United States with a mix of onsite and remote delivery. Some work genuinely requires boots on the ground. A lot of it does not, and treating every task as a site visit adds cost and schedule for no benefit. This article walks through what the team does, which parts typically happen onsite, which happen remotely, and what your team should expect to provide along the way.
What This Article Covers
- How Onsite and Remote Delivery Work Together
- InsightCM Deployment
- Ignition Screen Design and Integration
- AVEVA PI Integration
- KEPServer (KEPWare)
- 2D CAD Drawing Sets
- Fiber Optic Network Design and Installation
- Compliance and Site Access Expectations
- Onsite vs. Remote at a Glance
- Getting Started
How Onsite and Remote Delivery Work Together
A typical engagement splits into three phases. Discovery and design happen remotely: network topology review, tag inventory, panel locations, existing drawing review, and a scope document your team signs off on. Physical installation and commissioning happen onsite: mounting hardware, terminating sensors, pulling and testing fiber, verifying signals against the real machine. Configuration, screen development, historian mapping, and ongoing support happen remotely again, once secure connectivity is established.
That sequencing matters for budget. Site time is the expensive part of any integration project, and it is also the part that competes with your outage schedule and your escorts' availability. Pushing configuration work off-site shortens the window when Cutsforth personnel need to be inside the fence.
It also matters for the long tail. Screens need changes. Alarm thresholds get tuned after a few months of real data. New assets get added. Remote access means those changes take hours instead of waiting for the next trip.
InsightCM Deployment
InsightCM is the software layer that ties multiphysics measurement data into a single view of asset health. Deploying it involves more than installing an application.
Expect the team to work through server sizing and placement first, including whether InsightCM lives on a virtual machine in your environment, on dedicated hardware, or in a segmented monitoring network. Database configuration, storage planning for waveform data, and user account structure follow. Then comes device configuration: connecting condition monitoring systems to the server, setting acquisition triggers, and defining alarm rules that reflect how your machines actually behave rather than generic defaults.
Onsite work centers on hardware installation, sensor verification, and network connectivity. Server setup, asset hierarchy configuration, alarm tuning, and user training are commonly handled remotely. If you are bringing in sensors that did not come from Cutsforth, the third-party sensor integration path covers 4 to 20 mA analog, Modbus RTU, and Modbus TCP devices.
Ignition Screen Design and Integration
Ignition is widely deployed in power and process facilities as an HMI and SCADA layer, and many plants want condition monitoring data visible in the same environment operators already watch. That means building screens rather than adding another separate application for people to check.
The design conversation usually starts with who the audience is. An operator display and a reliability engineer's diagnostic screen are different products. Operators need a clear health indication and an escalation path. Reliability staff need trends, waveform access, and correlation across measurement types.
Expect to review wireframes before development begins, then iterate on a staged gateway. Screen development and testing happen almost entirely remotely. Final acceptance and operator walkthrough are worth doing onsite, because that is when you find out that the alarm banner is in the wrong place for the way your control room is laid out.
AVEVA PI Integration
If your plant runs AVEVA PI (formerly OSIsoft PI), it is likely the system of record for process data, and your reliability program benefits from having condition data and process data in the same place. Vibration without load context is much harder to interpret than vibration alongside megawatts, temperature, and excitation current.
Integration work covers tag naming conventions, mapping structure, write permissions, and update frequency. Two directions are possible: pushing condition monitoring data into PI so it appears alongside process tags, and reading PI tags into InsightCM so process context travels with the waveform. Recent InsightCM releases added support for reading digital PI tags, which widens what can be pulled in as context for a data event.
The main thing to prepare for is a tag naming decision. It seems minor during design and becomes very expensive to change once thousands of tags exist. Cutsforth will bring a recommendation, but your PI administrator should be in that conversation early.
KEPServer (KEPWare)
KEPServer handles protocol translation. Industrial plants accumulate equipment across decades and vendors, and the resulting mix of Modbus, OPC DA, OPC UA, EtherNet/IP, and proprietary drivers does not talk to itself. KEPServer sits in the middle and normalizes it.
Expect the team to inventory what needs to connect, select and configure drivers, set up OPC UA endpoints with appropriate security policies, and establish certificate trust between clients and servers. Anyone who has connected KEPServerEX to an OPC UA client knows the certificate exchange is where projects stall. Having someone who has done it repeatedly saves days.
This work is well suited to remote delivery once network access exists. Onsite involvement is usually limited to devices that need physical verification or that sit on networks with no remote path.
2D CAD Drawing Sets
Documentation is the part everyone agrees is important and nobody wants to own. It is also the part that costs you the most three years later, when a new technician is troubleshooting a monitoring circuit with no accurate loop diagram.
The Automation and Control team delivers drawing sets in CAD formats, and can also work from your existing drawings, updating them to reflect the services and hardware installed rather than handing you a parallel document set that conflicts with what you already maintain. That second option is usually the better one. Two competing sets of drawings is worse than one imperfect set.
Typical deliverables include panel layouts, wiring and termination diagrams, sensor location drawings, and network architecture diagrams. Field verification happens onsite during installation. Drafting and revision happen remotely.
Fiber Optic Network Design and Installation
Condition monitoring hardware often sits far from the control room, in environments with substantial electrical noise. Fiber solves both problems. It carries data over long runs without the distance limits of copper, and it is immune to the electromagnetic interference that surrounds generators, large motors, and switchgear.
Design covers path routing, single-mode versus multi-mode selection, fiber count including spare strands, enclosure and patch panel locations, and switch specification. Installation covers pulling, termination, splicing, and testing. Expect optical loss testing and documented test results as part of closeout, not as an optional add-on.
Plan spare capacity during design. The incremental cost of additional strands during the initial pull is small. The cost of a second pull two years later, during an outage, is not.
Compliance and Site Access Expectations
Working inside a power plant network is a security exercise before it is an engineering exercise. Cutsforth's team engages with the security standards your organization requires, including Critical Infrastructure Protection (CIP) requirements, background checks, multi-factor authentication, and specific device use requirements that restrict what hardware can connect to plant networks.
NIST's guidance on operational technology security, published as SP 800-82 Revision 3, is a useful reference for how remote access to monitoring systems should be architected. The short version is that remote access needs to be brokered, logged, and revocable, and monitoring systems should sit in a segmented zone rather than flat on the control network.
What this means practically: start the access paperwork early. Background checks and site access approvals routinely take longer than the technical work they gate. Bringing Cutsforth into that conversation during scoping rather than the week before mobilization keeps the schedule intact.
Onsite vs. Remote at a Glance
| Service | Typically Onsite | Typically Remote |
|---|---|---|
| InsightCM Deployment | Hardware installation, sensor verification, network connectivity | Server setup, asset hierarchy, alarm configuration, training |
| Ignition Screen Design | Operator walkthrough and final acceptance | Wireframing, screen development, testing, revisions |
| AVEVA PI Integration | Rarely required | Tag mapping, interface configuration, validation |
| KEPServer (KEPWare) | Devices with no remote network path | Driver configuration, OPC UA endpoints, certificate trust |
| 2D CAD Drawing Sets | Field verification and as-built markups | Drafting, revisions, final drawing package delivery |
| Fiber Optic Network | Pulling, termination, splicing, optical loss testing | Path design, component specification, documentation |
| Ongoing Support | Major hardware additions or expansions | Software updates, alarm tuning, screen changes, troubleshooting |
Getting Started
The most useful first conversation is a scoping call rather than a proposal request. Cutsforth's team will want to understand which assets you are trying to cover, what monitoring and control systems already exist, whether PI or Ignition are in play, what your network looks like between the equipment and the control room, and what your site access process requires. That call usually surfaces a shorter path than the one you had in mind.
If you are still building the internal case, the Reliability ROI Calculator translates downtime and maintenance spend into the financial terms leadership responds to. It is a reasonable place to start before you have a scope defined.
Cutsforth has more than 30 years in power generation and industrial services, with monitoring deployed across more than 1,200 facilities. The Automation and Control team brings that same experience to the integration work that makes monitoring data usable rather than just collected.
Talk with an expert about your project, or call 800-290-6458.
About the Author
John Pasquarette is a product and marketing leader with a long track record in industrial technology, engineering software, and IoT sensing. He has led product and marketing teams at companies including Cutsforth, National Instruments and Monolith, where his work has centered on helping engineers and manufacturers turn sensor data and analytics into better decisions. Based in Austin, Texas, he writes and speaks on condition monitoring, predictive maintenance, and the Industrial Internet of Things.