What are Operator Training Simulators?

2D and 3d Operator Training Simulators

Operator Training Simulators (OTS) include the simulation of the control system and the process. The control system displays are virtualized or emulated and connected to a dynamic process model. The industry standard control system user interface uses P&ID style graphics, faceplates, and tabular displays in a 2D format on a CRT screen.

Any 3D simulation of a plant is considered immersive in nature. Full scale nuclear simulators are immersive or ITS, generic simulations with a 3D model of a pump or heat exchanger is an ITS simulation, and crane/excavator simulators with 3D behavior are ITS.

Several companies are connecting OTS and ITS simulations. The console operator may train on a 2D display that matches the control system, and the field operator may train with an avatar navigating the virtual 3D process.

ARC treats the OTS and ITS parts separately, however these two classes are merging and may eventually be inseparable. In the long term, OTS will increasingly include 3D content of immersive simulations.

OTS and ITS software involves the simulation of the manufacturing process mechanics, geometry, chemistry, and physics. Typically, the console operator trains with 2D OTS software for refining, chemicals, pulp and paper, and pharmaceuticals. OTS software for robotics, discrete part manufacturing, and heavy equipment operations tend to use 3D user interfaces (ITS). The physical models used for the process can be high fidelity including models for nucleonics of various nuclear reactors, chemical reactor kinetics, equilibrium, distillation tray-to-tray, equipment models of pumps, compressors, piping, and heat exchangers. Such models can include extensive physical property behavior of complex mixtures of solids, liquids, and gases. In some cases low fidelity “tie back” models are used to check out and test control system software with an alternate use for operator training. A key function of the OTS software is how the control system algorithms and HMI are simulated. The HMI for control systems is essentially all 2D in nature and tends to mimic P&ID type drawings.

ITS software is 3D in nature and may involve 3D animation, virtual reality, avatars, and multiplayer game playing. Where games are modelled after reality, ITS simulations can be developed in the same way as games. The military is increasingly using ITS software for infantry and war planning, such as the “Dismounted Soldier Training System”. The level of immersion and the perception of reality can vary considerably depending on the HMI. 3D rendering can be done on a common computer screen with or without 3D glasses. The use of a game controller to navigate in 3D space is usually required, as a mouse is not optimized for this. Some ITS suppliers provide a “cave” where flat computer screens are floor to ceiling around a standing human using 3D glasses. There are also curved displays and domes for this purpose. There are dozens of new heads up display glasses. DAQRI can provide a heads up display helmet with many functions, although it is targeted more for augmented reality operation than training.

Operator training software suppliers come from a wide range of commercial origins. A few represent commercial spin-offs from user industries, partnerships, and acquisitions. Many firms focus on the training programs that accompany the technology. Others have developed simulation and optimization tools from steady state process design models. DCS and PLC suppliers have increasingly developed virtual versions of their control systems that can be software connected to process simulation models. Replicating the process automation solution and simulating the plant or machine behavior are both critical to training objectives. To be effective at training, the supplier must have domain expertise on the physical process, the instrumentation, and the control system.

Operator Training Simulator (OTS) Types

ARC refers to OTS software as the 2D process simulation that is integrated with the control system algorithms and HMI. The primary purpose of the OTS software included in this study is for console operator training for the process industries. Industries include the power generation, oil & gas, refining, chemicals & petrochemicals, metals & mining, pulp & paper, pharmaceutical & biotech, food & beverages and water & wastewater. The software OTS must be dynamic and market data includes the software and services only.

The OTS simulators must include the control system as part of the simulation and there are several options for how this is done. The trend is to use virtual control systems.

Virtual Control System

Virtual simulators use the actual code from both the control logic and the human machine interface (HMI). Typically, this is provided by the control system supplier along with an API (Application Programming Interface) to communicate to the process simulation.

The control logic (DCS/PLC/ESD) is supported to run in a Windows or Linux computer. Implementation may involve several virtual machines connected on a virtual Ethernet network all running in a single server that supports a remote desktop to a client PC for the operator control station. The I/O function blocks in the controller (like AIN or Analog Input) have a simulation toggle to disconnect from real I/O and receive a value from the process simulation. This results in precisely the same functionality as the plant’s actual controls. In addition, the simulator HMI screens use the actual DCS graphics, which are set up to run on one or more PCs. In both functionality and appearance, the HMI on the simulator is an exact reproduction of the actual graphics in the plant.

Hybrid System

Hybrid systems are combinations of the DCS HMI possibly with DCS consoles, annunciator panels, touch screens, control panels with buttons and lights, alarm horns, or other DCS specific functions. The control logic is virtual so it will run the same as the logic in a real controller. The operator interface uses the actual control vendor’s operator consoles, which are connected to the control logic. The virtual controller receives measurements from the process simulation, and control output are sent from the virtual controller to the process simulator possibly running in the same machine or a nearby virtual machine. This is very similar to the virtual system except the HMI may be equipped with some of the special consoles and accessories provided by the DCS supplier.

Fully Emulated System

Fully emulated systems use an emulation of both the control logic and the operator interface. Emulation is achieved by translating the plant control logic and graphics into the simulator environment, effectively reproducing the functional operation of the actual controls and the graphic interface of the HMI. This can be problematic as complex control system function blocks and the HMI may not be precisely emulated in the simulation computer

Full Hardware Stimulated System

Fully stimulated systems use the control vendor’s full hardware configuration for I/O, controllers, and HMI. In the early days of control systems every system was staged and tested with full hardware before it was installed in the plant. Most control system suppliers test components at the factory and ship the hardware to the plant to save time during installation and loop testing. Control engineering and HMI are typically implemented and tested on a virtual system. The full hardware stimulation arrangement would be the most expensive but it does provide the most realistic behaviors for timing and signal conditioning when the actual control system I/O is replicated in the simulator. In some arrangements, the hardwired I/O can be deleted and the process simulation connected to the DCS network with a software connection to the I/O function blocks in the controllers.

Operator Training Simulator Fidelity

Fidelity applies to the process models, the control system models, and the environmental simulation. The first rule of modeling is: All models are wrong, some are useful”. This applies to process, control, and environment.

Process Models

While steady state models are developed for a specific operating condition, dynamic simulation models operate over a broad range and dynamic behavior adds another level to model complexity. High fidelity simulations might compute variable within a few percent of the real variables. Maintaining model fidelity in the long term can be difficult as process changes or equipment wear like valve hysteresis or degradation of catalyst activity need to be adjusted in the process model. At the other extreme, some simulations merely strive for “tie back” behavior. When a pump is commanded to run, the “running” and “stopped” contacts are simulated so the pump status follows the pump command. Flow control loops may be simulated with simple “lag” function blocks (flow is proportional to valve position), and complex equipment may simply be set by the simulation operator. Low fidelity simulation can still provide great value when the focus is on control system training or testing. High process model fidelity is required when training is focusing on the process behavior.

Control System Models

As discussed elsewhere in this report, control system models can be virtual, emulated, stimulated, or hybrid combinations. In many training situations, accurately educating the operator about how the control system behaves is essential. It is increasingly common that control system suppliers support a virtual control system on a workstation, or a local or remote Cloud server. With virtual or hardwired stimulated systems, the control simulation is essentially a perfect match to the real control system.

Except for the most trivial control systems, emulation will suffer from inexact replication of the control system behavior. One might think that converting a PID in the control system that uses Proportional Band to an emulated PID that uses Gain is all there is to it, but that is not the case. Control system function blocks like PID have many unique behaviors with initialization, bad behaviors, output limiting, feedforward, ratio, and many other algorithm options. Accurate emulation of the whole set of DCS system function blocks is difficult to get perfect. This is why actual hardware controllers or virtual control systems supported by the control system vendor are commonly used. There is an advantage to emulation of control functions in the simulation environment to support special functions like “halt” non-real-time speeds, or playback.

Environmental Models

3D models of equipment or of a whole site may come from CAD drawings. These drawings may not reflect the final image of the equipment once it is painted with logo’s, safety warnings, and connected with other equipment. Site models made by using LIDAR imaging provide a wireframe model. Considerable “art work” is needed to create proper colors and textures.

Bentley has shown technology that can create 3D models from a series of photographs by stitching together overlapping edges. The game and media industries have many software tools for improving and rendering 3D images. While it is still not trivial to create the 3D files needed for a high fidelity simulation of a site and its equipment, it is cheaper and easier than ever before to do this, and is getting more cost effective every day. Training for abnormal situations, like fires, explosions, nuclear reactor meltdowns, will increasingly use immersive 3D training simulators.

Operator Training Simulator Functions

The following are some of the functions that are found in operator and immersive training simulation systems:

• Initialize simulator from preset operating conditions
• Use for abnormal condition simulation that operator responds to or modifies the process to correct the behavior
• Simulate plant shut-downs and start-ups
• Run, freeze, and restart operations from instructor's console
• Instrument, valve, pump, and other equipment failures
• Playback operator actions for reviewing
• Abnormal situation simulation
• Training and competency session reports and scorecards for operator proficiency
• Visualization can be multiplayer with avatars. The control room operator might use his traditional 2D control system HMI, while field operators are manipulating pumps, motors, and hand valves