Poorly designed and overly simplistic energy performance indicators (EnPIs) often drive energy savings at the expense of product yield or quality. However, a well-designed energy management information system (EMIS) can minimize energy cost without impacting production and, in some cases, can even enhance process performance.
Traditional energy monitoring applications mainly focus on improving energy-side key performance indicators for fired boilers and heaters efficiencies, energy intensity, utilities’ marginal cost, etc. These monitoring applications rely on inputs from various process measurement instruments, with temperature leading the way, to verify performance.
However, covering an expanded range of production parameters — including energy supply, demand and recovery, product quality and process yields — requires integration of the process with energy simulation, monitoring and optimization tools. This article shows how to overcome traditional barriers to energy saving by using rigorous process simulations to monitor performance and determine optimum operating targets for improving both energy and process performance.
The Energy Opportunity
Energy is the largest controllable operating cost at most process plants. A typical refinery or petrochemical plant may spend $200–300 million/y on energy — so cutting just 3% in energy cost can save $6–$9 million/y. Such energy savings always result in direct bottom-line benefits, unlike adding capacity or changing product mix, which depend on anticipated market conditions.
Energy production and distribution systems often constrain processes. For example, a process compressor can be limited by its turbine drive’s capacity and efficiency, so steam and condenser operating conditions or degradation of the turbine can mean the drive reaches its limit before the compressor does. In another example, the amount of heat a process furnace is able to deliver can restrict unit throughput. Energy-related bottlenecks often curb throughput of high-margin processes by 2–3%.
One challenge is understanding the amount of potential energy improvement. Plants typically compare themselves against their peers. However, this comparison only is meaningful if the leaders are highly efficient.
An alternative approach is to compare energy use against a thermodynamically and economically achievable minimum. Our company has developed an energy metric called the Best Technology (BT) index. The target BT index is calculated based on an optimized process configuration including reactor conditions, number of distillation column trays, etc., as well as pinch analysis for heat recovery and R-curve analysis for utility delivery. This enables the specification of all equipment for maximum efficiency.
Pinch analysis is a methodology for reducing energy consumption of processes by calculating thermodynamically feasible energy targets. R-curve analysis determines the hypothetical ideal utility system and fuel utilization for power and steam generation.
Repeating these optimization calculations for a range of feedstocks, operating severities and product yields determines a relationship between optimum energy use and process performance. The optimum target energy benchmark is defined as 100.
The actual BT index is calculated as the ratio of actual energy use divided by the target, in %. For example, if the plant is using twice as much energy as the benchmark, then its BT index is 200%. This index basically compares current energy use against that of the best available technology in the market.
Even relatively efficient plants typically use significantly more energy than the BT benchmark. Figure 1 shows a trend of the BT index for several hundred sites arranged in descending order along the x axis. In the refining and upstream industries, even the best performers (right-hand end of the scale) have a BT index well above 100%. There have been many instances of top performers saving 10–15% of energy, worth $20–30 million/y. Similar percentage reductions for CO2 emissions also are achievable.
Most EMIS software packages focus only on the energy supply side (for example, the efficiency of production of steam and power for use in the process), so their EnPIs don’t reflect the impact of feedstock effects or process yield. For instance, if energy consumption increases, they can’t indicate whether this stems from inefficiency, lower quality feedstock or the demands of higher quality products. These software packages may monitor equipment performance but often miss the chance to switch an item of equipment off when its output isn’t needed to support production.