To help quantify the performance gains and savings that the Open Process Automation Standard (O-PAS) can deliver, the non-profit Coalition for Open Process Automation (COPA) reported on a recent, front-end engineering design (FEED) study by one of its clients, which evaluated the initial and long-term costs of operating three traditional distributed control systems (DCS) compared to an O-PAS distributed control node (DCN) system. The client is a multinational, agricultural products company that reports it’s an active, non-OPAF member, and tested O-PAS at one of its continuous process plants with 14,000 I/O to conduct its head-to-head cost-comparison study.
In the first part of the study, COPA-member Wood served as engineering procurement contractor, and calculated the initial cost and 25-year total cost of ownership (TCO) of O-PAS versus the DCSs. In the second part, the client also calculated the 25-year TCO.
Using a scalable, COPA-based DCN, the contemplated, greenfield installation excluded field device, construction and cabling costs, and included: one-time costs for control system hardware and software, engineering, configuration and acceptance testing; ongoing costs for owner and contractor labor for control system operation and maintenance; and recurring costs for lost production from plant downtime due to control system maintenance. It also includes costs of frequency changes, such as Windows or Linux patching at one month; firmware upgrades at one year; office network replacement at five years; HMI replacement at 10 years; industrial network placement at 12 years; control/compute replacement at 20 years; and I/O placement at 25 years.
The study found that:
- Initial costs for system hardware and software were 52% less with COPA’s DCN compared to the first DCS, and 10% less for the total project, according to the EPC’s calculations.
- TCO over 25 years was 47% less with the DCN than with the first DCS, according to the EPC’s calculations; and
- TCO over 25 years was 60-70% less with the DCN compared to each of the three DCSs, according to the agricultural firm’s own calculations.
One reason an open-computing O-PAS DCN can generate savings is that it typically has a big advantage in computing power and speed compared to a traditional DCS, while it also has software redundancy with no incremental hardware costs. For instance, while two of the DCSs that Wood and the agricultural client studied a few months ago can perform 760 million instructions per second (MIPS) or 855 MIPS for $20,000-$50,000 per device, a DCN such as COPA-member ASRock’s industrial PC (IPC) has 10 microprocessor cores that can perform 30,000 MIPS for about $2,000 per device, which lets it run more sophisticated algorithms and provide results more efficiently at far less cost.
“Usually, we can only run MPC at the top of the computing stack once every 30 seconds or so, which can make it hard to use results. However, an O-PAS DCN can let MPC run right in it, which also means less bandwidth use and latency,” says Don Bartusiak, former co-chair of OPAF and president of COPA-member Collaborative Systems Integration Inc. “Driving more computing power down from Level 3 to a DCN at Level 1 can allow users to solve other types of problems that will add value.
Bartusiak reports that more computing power lets users employ artificial intelligence (AI) or neural networks to better estimate compositions for distillation towers in a more timely way than using a gas chromatograph, or use deep-reinforcement learning to continually optimize PID loop tuning. It can also enable software redundancy with software pairs that can pick up if one fails, or allow orchestration to achieve high-availability by implementing redundancy in a dynamic, software-define way.
“Eliminating the need for physical pairs for additional computing hardware also means fewer hardware backups are needed, which allows spare capacity in the system to be used for more software-based functions,” adds Bartusiak. “The computing power in multiple DCNs can provide redundancy without adding hardware costs. Plus, this is reusable engineering software that can greatly improve ease of use. However, the biggest savings will come from decoupling I/O from computing functions because users will be able to upgrade at the compute level without touching wires and I/O.”
To deliver these high-speed capabilities, ASRock and other upcoming DCNs must be able to reach out, network, communicate and collaborate with the other devices in their overall operations. O-PAS defines and accomplishes this task for DCNs by establishing and verifying device profiles in its O-PAS Systems Management (OSM)-003 conformance requirements section. Profiles are usually a series of “shall” statements that determine and assign functions like connectivity, networking and security. This verification process is based on the non-profit Distributed Management Taskforce’s Redfish Agent Standard for IT and cloud-computing services, and it’s designed to allow device access and vendor-agnostic visibility of system resources. This enables system orchestration, which is automated applications, requirements and lifecycle management of devices, but with the added benefits of with zero downtime and minimal required IT skills.
