Most process plants consider steam as an indispensable means of delivering energy. After all, it offers many performance advantages including low toxicity, ease of transportability, high latent heat and low cost of production. Because most of the energy in steam is stored as latent heat, large quantities of heat can be transferred efficiently at constant temperature.
Typical steam systems encompass multiple pressure levels connected to a number of steam producers and steam users or consumers spread across a site. As economies of scale drive operating companies to build ever larger and more integrated facilities, the design of the shared steam utility system becomes extremely critical to their operation. The steam headers often run throughout the complex, tying together myriad units. This creates a highly non-linear control and operability challenge.
It's essential to ensure that steam can be provided to all reaches of the facility without interruption and that the system can be controlled in the event of upsets to maintain stable operation. Improper controls could lead to loss of the entire steam system, trip or damage of critical equipment, off-specification products and, in the worst case, loss of the entire steam system and shutdown of the complete facility. Normally, such design deficiencies become apparent only after an incident — this could be costly or potentially disastrous.
Further, with ever-increasing energy costs, better design, control and operation of the steam system can directly impact the entire facility's overall efficiency, translating into substantial operational savings.
Traditional steam hydraulic analyses assess demand and production issues at different steady-state operating conditions. Such analyses can't predict the steam system response through multiple headers all across the complex during process upsets.
Understanding the response through dynamic transients and ensuring the steam system can handle all expected events without jeopardizing the availability of the facility becomes a critical aspect of the process and controls design of such systems.
This article takes a look at how dynamic simulation can assist steam system design and offers up some tips for staying out of "hot water."
AN IMPORTANT TOOL
Dynamic simulation is a "best available technology" that can be used to evaluate the "as designed" process and control strategy to maximize the likelihood that it can provide stable and uninterrupted operation following steam system or process upsets.
A typical dynamic simulation of the steam system involves building a rigorous first-principles model that includes:
• boilers;
• steam turbine generators and drivers;
• multiple pressure headers;
• pressure letdown stations;
• steam consumers; and
• regulatory and plant master control.
Today commercially available software packages such as DYNSIM from Invensys allow steam system models to be built in a fraction of the time of older programming languages or software platforms (Figure 1).
The model is built and the controls are configured to maintain the steam system at the normal design operating point(s). The model encompasses all regulatory controls, including those specifically designed to manage expected transients resulting from steam-system or process upsets.
The high fidelity model can simulate many steam-system or process upset scenarios in a matter of just days — to predict the system response following such events in a safe and controlled environment on the computer. The model then can be used to determine how best to correct any issues identified during the upset scenarios.
TIPS AND STRATEGIES
While dynamic simulation has become more prevalent due to software and processor advancements, it involves far more than simply entering numbers into a form. Experience has shown that certain considerations and pre-planning strategies significantly contribute to the success of steam-system simulation projects. Here are some tips:
Make conservative assumptions. This is one of the most critical aspects of design. It's inevitable that the model won't capture every possible nuance or feature of the process, so the model's response won't fully replicate actual system response. However, if the model is designed to the highest possible rigor and all assumptions and modeling approaches err on the side of safety and over-design, you can have confidence in the results.
For example, Gandhi et al. [1] discuss the modeling of steam systems (specifically boilers following a trip) and the phenomenon of self-boiling in which residual heat in the boiler continues to generate steam long after fuel is cut off. While it might be possible to rigorously model the boiler to the level of detail to capture the self-boiling phenomenon, it may be more prudent instead to take a more conservative approach — assuming steam generation stops shortly after fuel is cut off. A control system that can handle a rapid loss of steam certainly can deal with the situation where the steam supply decays more slowly.
On the other side of the coin is modeling the ramp-up of boilers when more steam is needed. The boiler manufacturer will supply the design maximum rate of change of steam production up to the maximum continuous rating. The vendor may give a 20% per minute ramp but what if it's actually only 10% or 15% due to unforeseen issues. The dynamic simulation platform provides a perfect environment to run multiple cases to test the sensitivity to key parameters.
Employ strict quality-assurance procedures. The accuracy of simulation results depends upon many factors, including the modeling approach, assumptions, data mining and data input. Experts follow strict procedures when executing a project to ensure the model is built to the highest possible standard and model inputs are correct. It's crucial to establish quality-assurance procedures that will certain results obtained are meaningful and trustworthy.
The main focus should be on checking the data input into the model. Discuss assumptions made and confirm they're conservative enough that the results won't compromise any objectives of the study. For instance, using a larger steam header volume than actual in the model could yield a slower response than actual; this could lead to inaccurate results and conclusions for the design of pressure controls.
In addition, have experts from operations review scenarios that are tested on the model to ensure the worst case is considered.
Such a quality-assurance process guarantees the model developed includes all the right inputs and assumptions, making the results more reliable and credible.
With a high-quality steam system model, engineering and operating companies can begin to reap the benefits of dynamic simulation in different aspects and phases of the design process.
Validate steam-network piping design. The piping network often is designed for the flows and pressure profile at steady-state conditions. In the event of process upsets and the transients that may follow, these parameters undergo rapid changes that normal hydraulic analysis can't discern.
Some pipes within acceptable limits at normal conditions could exceed design limits during transients and become potential bottlenecks to the steady operation or startup of the facility. Identifying such bottlenecks during startup, commissioning or after an incident could lead to expensive field changes that impact the project schedule for new plants or operation of an existing facility.
A dynamic simulation analysis of the steam system helps precisely understand transients in the system. Simulating upsets enables monitoring flows and pressure drops across pipe segments as a function of time, to identify violations of design criteria during the transients. The greatest benefit from this type of analysis occurs when it's performed closely with the engineering design of the system. At that point, incorporating necessary design corrections incurs the lowest possible cost and impact on schedule.
Confirm steam system controls. Dynamic simulation also can be used to evaluate the proposed control strategy around the integrated steam system. It can help get the control system right the first time, thereby saving valuable time during commissioning, helping ensure stable operation during day-to-day operations and keeping the system up and healthy during some of the worst-case scenarios the facility could experience.
An upset, like loss of a boiler, has the potential to bring down the entire steam system, causing shutdown of critical process units. Because a dynamic simulation model incorporates all the controls, analysis can determine if the as-built controls can maintain stable operation after an upset. The model allows easy configuration and testing of control alternatives that might improve steam-system response. Feedforward signals to boiler controls, low/high selector clamps on letdown stations, priority settings on steam headers, set-point staggering across the facility on various control loops are some of the important handles that can be quickly changed and fine-tuned using a dynamic simulation analysis. These parameters can prevent nuisance trips and shutdowns and can accelerate startup.
Identify steam load-shedding strategies. A critical outage of major steam producers for scheduled maintenance or due to an unforeseen trip requires adjusting steam demand to balance supply and demand across the complex. If backup boilers can't make up the difference or are slow to respond to the upset, it will become necessary, for example, to switch from steam-driven turbines to electric drivers (if available) or to identify which less-critical units should be taken offline and for how long to protect the more-critical equipment and units.
When transient demand exceeds transient production, as in the case of multiple boiler trips, a steam shedding strategy must be initiated quickly to counter the upset before steam networks reach unacceptable pressures. Developing a steam shedding plan that could be implemented during a major upset is critical to maintaining the availability and un-interrupted operation of the steam system.
Dynamic simulation can be a great help with this evaluation as it can be used to test and evaluate critical shed lists and to develop a strategy — prior to startup and operation — that least impacts the economic profitability of the overall complex. It allows analysis of either reducing steam consumption or dropping steam users outright based on priority and criticality. Both feedback pressure-driven and feedforward event-based shedding strategies can be easily configured and tested.
KEEP OUT OF HOT WATER
Dynamic simulation quickly is becoming an accepted technology for performing in-depth steam system analyses that can't otherwise be done except by trial and error in the plant. Engineering companies can benefit greatly from performing such analyses by following the tips described here and other best practices as early in the process lifecycle as possible.
Moreover, the simulation software platform models become assets within the company and can be re-used beyond the design environment to support plant commissioning and for the development of operator training systems.
Avoid Common Design Errors
Dynamic simulation can aid in a number of areas of steam system design, including:
Properly sizing lines. One of the common errors encountered in steam system design is incorrectly sized distribution piping. Undersized lines have higher velocities and pressure drops, leading to insufficient flow and pressure of steam to users. Undersized lines also increase the risk of erosion, noise and hammering. On the other hand, oversized lines are expensive and cause higher heat losses, impacting the quality of steam. In addition, flows through steam pipes can undergo drastic changes. Understanding this phenomenon through simulation is crucial for accurately estimating and verifying line sizes.
Getting the system control loop right. Feedback loops alone might not suffice to control the steam system through a wide range of upsets. Scenarios where steam supply exceeds demand can be handled by disposing of excess steam for a short while until feedback loops bring the system under control. However, when there's a sudden shortage of steam, the feedback control actions might be too late. Appropriate feedforward control actions must be initiated before the shortage affects operation of the facility. Understanding the extremely non-linear characteristics of a steam system by simulation is critical for successful design of these feedforward controls.
Setting correct priorities in steam shedding. Situations where there's likely to be a severe shortage of steam require an emergency steam-shedding plan to avoid a full-scale shutdown. But which units should be shed and in which order? Such situations require a carefully designed strategy that prioritizes shedding of steam users based on the impact on the overall operation. As with designing feedforward controls, a simulation model can be used to evaluate different steam-shedding strategies in a cost effective and safe manner, thereby ensuring the best possible emergency shedding strategy is determined and deployed in the master controller.
IAN WILLETTS is a director of Invensys Operations Management, Carlsbad, Calif. ABHILASH NAIR is a principal consultant for Invensys Operations Management in Carlsbad. CHARLES REWOLDT is an application engineer for Invensys Operations Management in Carlsbad. E-mail them at [email protected], [email protected] and [email protected].
Reference
1. Gandhi, S.L., Graham, J., Duffield, M.A., and Cortes, R.M., "Dynamic Simulation Analyzes Expanded Refinery Steam System," p. 3, Hydrocarbon Processing (Nov. 1995).
