Predictability in today’s capital-intensive markets is a challenge. While capital spend is up 400%, revenues are only up 2%. Project cost overruns across all industrial markets are one of the areas that are out of control and contributing to this dynamic.
One of the largest companies in the world recently stated their project execution nightmares as:
Project cost overruns,
Missed schedules,
Missed production dates,
All causing unacceptable and unpredictable risk.
Honeywell has conducted extensive voice of the customer (VOC) analysis on this topic and has found that these same issues and dynamics are experienced by companies of all sizes, all over the world. Capex planning, returns on invested capital and cash flow are all impacted negatively by these inefficiencies. Significant work has been done to improve project efficiencies through the use of integrated database tools and workflow, but this has not really addressed some of the fundamental issues. Fine-tuning the project execution process can result in incremental changes, but does not really move the needle. In fact, processes used to implement plant automation and safety system projects have not changed much in the last 40 years.
When it comes to large greenfield projects in the industrial world, modular construction tec hniques are driving the need for early hardware shipments and reduced time to integrate components. In addition, global sourcing, geographically distributed plant design and engineering, and remote locations of construction sites make the project lifecycle increasingly dynamic by nature. Globally delivered projects require coordination across multiple offices and locations. Mega-projects often involve m ultiple engineering procurement and construction (EPC) companies around the globe with different project scope. Resource constraints may involve executing multiple projects or sites at the same time. Resource development in emerging geographies also means more projects in locations with t he fewest resources.
After decades of executing automation projects in the traditional way, it is time for new thinking. Why is it necessary to order physical systems at the beginning of the project? Why must groups of people travel across the globe to remote locations? Why force hardware freezes prematurely? And why should design be coupled to hardware? Now, more than ever, a paradigm shift in project execution methodologies is needed to increase capital efficiency and ensure predictability in cost and schedule on automation projects worldwide.
Traditional project implementation
Conventional project engineering processes progress through a series of activities that continuously build upon each other to move the design from abstract ideas to validated engineering designs. Through the progression of these activities, decisions a re made which enable subsequent activities to move forward based on the final design (See Fig. 1).
Fig. 1: Engineers rely o n a progression of activities to design the automation facility.
In the process industries, the plant automation system is a collaboratively engineered system in that its design typically is not known at the time of procurement; instead, the supplier develops the system from multiple requirement inputs derived from various engineering activities.
Due to the sequential nature of the traditional engineering approach for automation systems, there is only one path to project completion. Tasks are executed in fixed sequence to meet install dates, and outputs from one task are required for the following task. This single path means that any delay or issue arising in one of the activities will directly affect the delivery of the subsequent tasks. (See Fig. 2).
Fig. 2: Traditional automatio n project workflow.
The ultimate efficiency for this model is to execute each one of these tasks in the timeliest manner, and only execute it once. In a real world situation, this becomes a very challenging endeavor. Projects are often required to be finished on tighter and tighter schedules, and changes in a project’s scope or definition often occur in its later stages. These late changes may mean that a design will need to be changed, the build and/or configuration modified, and potentially, the testing of the component repeated.
Evolving project strategies
During new construction – particularly on mega-projects – it is crucial to minimise risk, increase flexibility, shorten schedules and, most importantly, keep automation systems off the critical path. This requires the elimination of non value-added processes, which can be considered aste. Typical examples include repetitive tasks, rework and redundant tasks
The demands placed on automation system suppliers to accommodate evolving project execution models have required reassessment of traditional delivery models. A paradigm shift in project implementation strategies means the workflow on projects can now be very different. New approaches to project execution can optimise processes and eliminate non -value-added activities that were once highly problematic. These techniques are intended to drive effectiveness — not efficiency — since to efficiently do what is not required is not effective.
Changes in delivery models
In contrast to the traditional automation project workflow described above, configuration of the automation system software to meet the functional requirements best favours a more iterative approach. With this technique, the system is built up as functional units. The units themselves have no dependence upon the hardware design state or the completion of other non-related units.
Separation of the physical and functional aspects of the control system into independent hardware and software design activities enables both of these tasks to be performed in parallel. This breaking of the dependence of the software configuration from the hardware delivery also allows configuration activities to occur out of the traditional order of tasks.
In addition, configuration activities can be started much earlier in the project schedule, as they are not reliant on the physical system being designed. This independence leads to the creation of two separate execution paths, which can be managed to meet the project’s deliverables with greater flexibility than the traditional model (See Fig. 3).
Fig. 3: Optimised project workflow.
Transformation of workflow
Changes in project implementation strategies now support the traditional (i.e., define, design, manufacture and install) workflow for th e physical hardware components and an independent workflow (i.e., define, configure, test) for the functional software. In the case of functional development, there is no penalty for breaking the workflow into modular functional packages. For example, a utilit ies process unit could be functionally completed prior to starting another production unit that may have a less complete definition. This enables the project team to be far more agile in supporting the fragmented data inputs that result from parallel upstream design activities.
In the evolution of process automation, the functionality of control systems has moved steadily away from dependency on hardware to software configuration. The functionality can be applied to multi-purpose hardware supporting a greater variety of functionality with fewer, more versatile standardised components. In this sense, the software configuration supports control system functionality while the hardware supports physical requirements, such as input and output (I/O) connections that are truly universal and can be changed remotely.
Advent of new technology
Thanks to abstraction technologies offered by Honeywell Process Solutions, including universal I/O (UIO), virtualisation and cloud engineering (virtual engineering platform), industrial firms can transform the way in which plant automation projects are executed. These technologies are specifically designed to decouple physical design from functional design. They support key project benefits such as: late binding of system configuration data, flexible hardware procurement, improved agility and flexibility, and enhanced design options. Honeywell’s solution – lean execution of automation projects (LEAP) – can have a significant impact on large capital mega-project implementation, taking millions of dollars off the total installed cost of large control system projects.
With the latest technology developments, it is now possible to achieve full abstraction of the control system infrastructure. Supervisory Control Level 2 nodes are abstracted via virtualisation, while control level 1 and I/O can be fully simulated in a server environment and abstracted by the universal I/O. This overall abstraction from the server to the I/O is fostering the new LEAP approach, which brings greater versatility to the traditional model of an engineered automation system.
Universal I/O
In the process of physical design for control systems, the objective is to start the design task as late as possible. Removing configuration activities from the critical path is one step towards that goal. Another step is to reduce the time taken for design by simplification of the design process itself.
The development of universal channel technology has completely liberated field I/O, as well as control cabinets, from channel-type dependency. This more standardised, multifunctional solution permits late additions and modifications to I/O schedules with no more than a soft configuration change — potentially saving weeks of schedule delay when making late-stage design changes.
Late binding: Through the use of universal I/O, late data binding can occur well into the construction phase of a project without the huge impact late instrumentation design changes would typically bring to conventional designs. Users no longer require different modules for each I/O type, and are able to reconfigure any I/O channel without having to go into the field to make physical changes. Additionally, functional design can be separated from the physical design. This allows for parallel work operations (i.e., functional and physical design can happen independently and concurrently). In situations where there is limited physical access to the control system during commissioning or offshore operations, the ability to make changes from a remotely connected engineering station can prove invaluable.
Flexible hardware procurement: Since universal channel technology has ushered in a more standardized design approach based on the use of a single, universal I/O module, standard cabinet designs can be manufactured later in the project schedule. Plus, fewer spares will need to be available as part of a project.
Improved agility and flexibility: With the universal I/O solution, project teams can start work sooner — they just require a total I/O count and do not need to worry about the I/O mix. Universal I/O modules reduce equipment requirements and footprint, and can be quickly configured for multiple channel types, ensuring utmost flexibility in system design. This concept enables multiple remote locations to be controlled out of a single centralised unit, with each channel of I/O individually software-configured either as analog input (AI), analog output (AO), digital input (DI) or digital output (DO).
Enhanced design options: universal channel technology makes it possible to utilize remote cabinet designs, with corresponding savings in equipment space, power, cooling and weight requirements. It eliminates wasted I/O space and enables reductions in both installation and operational costs since users are no longer concerned about having sufficient modules for AI, AO, DI or DO configuration. The I/O connection can easily be configured, and reconfigured, at any point.
The application of standardised universal cabinet designs to control system projects also allows for efficiency improvements in many areas in the design process and I/O assignment. In addition to these improvements, universal vabinets bring cost reductions and schedule gains in other parts of the project.
With universal cabinets, any field signal can be connected to any I/O channel. The control system can be sized purely on the estimated total I/O count for the overall process or process units. Taking these benefits into the design process, there is no longer a requirement to custom-design each cabinet based on the specific I/O mix for that location. A universal vabinet containing nothing but Universal I/O can be designed once and installed in any location for the project (See Fig. 4).
Fig. 4: Universal channel techno logy allows I/O cabinets to be standardised, because any field signal can be connected to any I/O channel.
For instance, a typical mid-sized process unit might require 10 I/O cabinets, with each cabinet containing around 500 input and output signals. To build these cabinets using conventional I/O, it is assumed each cabinet will need to have the appropriate mix of I/O arranged internally, as well as custom power and heat calculations. The cabinet layouts will have to be sent to a designer to produce internal arrangement drawings for internal and client approvals. After manufacture, the cabinets will be inspected and validated against the custom arrangements. Specific custom factory inspection and test records will need to be produced, with all of the cabinets assembled together and powered-up and tested as part of the customer’s acceptance test.
If the project had chosen universal vabinets, many of the aforementioned design activities could have been eliminated. Since all of the cabinets would be the same, there is no longer a need to produce custom layouts and drawings. Plus, no additional approvals would be necessary, since the quantity of the cabinets would be decided at the definition stage of the project. Manufacture and test requirements would also have been met by standard factory processes. Optionally, the end user could have chosen to witness the testing, with the cabinets then shipped directly to the site without the need for further power-on and staging.
If at a later point in the project the definition changed and more I/O was needed, this requirement could be easily and quickly accommodated by spare capacity within the cabinets. Or, additional standard cabinets could be manufactured with minimal lead-time and shipped directly to the site without impacting other project activities.
Virtualisation
Plant managers seek to reduce the amount of computer hardware in their facilities and their total cost of ownership (TCO), but they must do it without compromising existing safety, reliability and production. Simply put, virtualization abstracts operating systems and applications from the underlying physical infrastructure by representing the hardware as virtual devices. Virtualidation allows a single server to simultaneously run multiple operating systems and applications. It does this while insulating these virtual machines (VMs) from the underlying hardware and also from each other.
Fig. 5: With virtualisation, it is now possible to achieve full abstraction of the control system infrastructure.
Virtualisation is now proving to be an important productivity and efficiency tool for all types of industrial operations. Targeted virtualisation solutions simplify and decrease requirements for plant hardware, while enabling greater reliability and availability of process control systems.
Late binding: Virtualisation makes a significant contribution to design independence on automation projects. Traditional methods of deploying process control systems have involved a tight coupling between control functions and the instruments connected to the process. By employing virtualisation solutions, development of supervisory layer VMs can occur independently of the physical virtual infrastructure design. Virtual machines can be installed on the final physical virtual infrastructure late in the project cycle, and development of these machines can occur at the same time as the physical virtual infrastructure design. All of Honeywell’s simulators can also run as VMs, enabling implementation of changes to occur independently of hardware installation on site.
Flexible hardware procurement: When projects are provisioned with virtual machines rather than physical servers, virtual infrastructure purchases can be deferred until later in the project. The actual hardware is not procured until after final testing so that users can buy the most current technology, thereby avoiding a costly hardware refresh during project execution. Given that hardware is shared rather than dedicated, an estimated number of host servers can be used to meet the needs of a system without having to identify the exact server and operator station requirements. Virtualisation allows for a single, standardized workstation and server design that can accommodate a wide variety of topologies in a similar manner to universal cabinet designs.
Improved agility and flexibility: Plants also achieve greater agility and flexibility with virtualisation technology. Virtualisation allows project engineers to add new virtual machines without requiring new hardware (within limits). As such, they can start work sooner without worrying about a precise VM count. Even when hardware has been purchased, if additional nodes are required, they can be handled on the existing hardware assuming sufficient capacity. And using virtualized templates, new nodes can be spun up effortlessly without having to do full installs.
Enhanced design options: Virtualisation, coupled with the deployment of blade server technology, allows for a standardised, yet modular supervisory design that can handle a wide range of system design scenarios on new construction projects. Additionally, the technology can reduce the physical footprint of hardware components in the plant server room, with corresponding savings in space, power, cooling and weight. Facility savings are both direct (consumed by the hardware) and indirect (ancillary service reductions). Running these virtualisation solutions on blade hardware then provides the ultimate synergy of resource utilisation provided by virtualisation with the hardware density provided by blade technology.
Additionally, virtualised servers with a thin client operator interface are the best option for many applications commonly found in the plant control room. Like the server room, footprint can be optimised in terms of equipment space, cooling and noise. This is achieved through the thin client’s compact form and lower power usage. Additionally, thin clients have the flexibility to be geographically separated from the facility, using reliable redundant network connectivity.
Cloud engineering
Today, more and more engineering and industrial companies manage their projects across multiple locations. A lot of project teams aren’t defined by geographical proximity anymore. Instead, engineering professionals work as a ―human cloud that is made possible thanks to the rapid development of new cloud computing technologies. For companies that have global development and manufacturing operations, the ability to share data without being limited by time or place helps their various locations collaborate more closely. It also allows engineers to concentrate on their work rather than how, when, and where their jobs are executed.
Late binding: Cloud engineering solutions enable automation work around the world to be engineered in the cloud in a manner totally decoupled from the physical infrastructure. The VMs or databases can then be bound late in the project cycle with the physical hardware.
Flexible hardware procurement: One basic advantage of the cloud engineering approach is the freedom to use the automation supplier’s infrastructure during the project engineering process — not the customer’s. This allows the project to procure the project servers as late as possible in the schedule, ensuring that the technology delivered with the system uses the latest hardware available.
Improved ability and flexibility: With a virtual engineering platform, subject matter experts (SMEs) have fewer physical constraints and can start work on the project without waiting for servers or other components. As all of the control functions have been virtualized, remote testing is possible. This can be done through allowing more iterative remote inspection during the detail design phase of a project, reducing the need to test these areas during a factory acceptance test (FAT) or, alternatively, a traditional FAT could be conducted where large portions are performed remotely. Either way, large reductions in travel expenses are achieved on the order of 50-70% for testing. Once the virtual FAT has been completed, the virtual machines are moved from virtual staging resources to the final servers for the site acceptance test (SAT). As the machines have been virtualised, nothing needs to be reloaded or reinstalled when they are transitioned to the final hardware (See Fig. 6).
Fig. 6: Virtualised engineering environments in the cloud allow engineers to work together on projects from locations around the world.
Enhanced design options: Cloud engineering can deliver the functionality required by industrial firms that want to transform the way in which they operate to achieve a competitive advantage. Companies are running applications in the cloud to lower costs, deploy more quickly, and simplify infrastructure management. The ability to detect and react quickly to issues reduces waste and elapsed time, which reduces costs and improves time to market.
The combination of virtual engineering with full virtual target system deployment yields the highest possible benefits in terms of engineering design efficiency, quick deployment, late change management, flexible hardware procurement, and footprint for automation projects.
Advantages for industrial projects
When applying optimised project workflows enabled by Honeywell’s innovative control system technologies, a number of tasks that previously occurred in a system-staging environment can now be pushed out to the construction phase to enable earlier on-site delivery of hardware. With Universal I/O, users can now order fully standard universal cabinets from the factory in whatever quantities required, and execute a project without assigning I/O to a channel until commissioning. With virtualisation and product simulators, control strategies can be developed and tested prior to the final design, and users can develop those strategies in the cloud and allow the project engineer never to leave home.
Applied together, these technologies enable a new project execution methodology that revolutionises project execution. By separating physical from functional design, this methodology allows parallel workflows, applies standardized designs, and enables engineering to be done from anywhere in the world. Overall, a reduction in main automation contractor (MAC) costs could result in 30% capital savings and optimising scheduling by 25%.
For an integrated main automation contractor project of approximately $50-million with 40,000 I/O, this methodology could potentially deliver the following benefits:
66% reduction in defect opportunities and marshalling labor due to elimination of multiple terminations per loop
90% reduction in marshalling cabinets from 120 to less than 10 due to moving the I/O to the field
50% reduction in servers and network cabinets through server consolidation
60% reduction in remote instrument enclosures (RIEs)/local equipment rooms (LERs)/field auxiliary rooms (FARs) from 7 to 3 due to decreased footprint of automation equipment
Shipment of control system I/O hardware up to six months earlier to optimise field construction schedules due to standardised cabinet designs
50% reduction in travel expenses through virtualised engineering development and virtual acceptance testing
Finally, reduction in total MAC costs of about $16-million (30%) through the reductions stated above
Additionally, a large capital project could realise four to five times that amount in total savings when the impacts of LEAP are applied across an entire project. The savings would come from reduced capital costs, improved operational readiness and early production, installation and commissioning, engineering time, document control and review, schedule optimisation, travel time, and the typical delays and change orders incurred on most projects due to late design changes. Cutting back on rework, buildings needed and footprint can have far-reaching benefits across the project. For example, the reduction in design engineering documentation and drawings alone through elimination of marshalling designs and rework could dramatically affect the engineering hours on a major project. Reduced footprint of equipment through standardisation not only saves space, but reduces the number of buildings/enclosures dramatically – and the associated shipping and logistics, cooling and power requirements that support them.
Benefits to an engineering procurement construction (EPC) contractor:
Reduce construction costs
Reduce risk from late changes
Enables effective deployment of resources
Project delivery is more flexible
Benefits to the end user company:
Automation costs are predictable
Project schedule risk is reduced
CAPEX is more efficient
Earlier production dates
Conclusion
The benefits of optimised project implementation strategies coupled with the latest automation technology advancements, can be realised at both ends of the project schedule and represent a true paradigm shift. This approach may be applicable individually to a single project, or applied broadly to gain the maximum project flexibility.
LEAP utilises three innovative technologies – universal I/O, virtualisation and cloud engineering – to enable parallel workflows, standardised designs, and more efficient remote engineering. The resulting optimisation in scheduling, equipment footprint, labour and engineering time can be significant.
Across an entire construction project, the cost benefits can be multiplied four to five times, allowing both EPCs and end users to realise dramatic improvement in project execution. This new project implementation approach is essential to ensuring control systems pose the least risk to project schedule, cost and risk and allow for safe, efficient and reliable plant startup and operation.
Contact Boni Magudulela, Honeywell, Tel 011 695-8033, boni.magudulela@honeywell.com
The post New solutions for optimising delivery of automation projects appeared first on EE Publishers.