Inside a demand-response system

Many industrial users of energy think demand-response schemes made possible by smart grid technology are a great idea -- for somebody else. When you operate time-critical industrial processes like furnaces or conveyors, it is hard to envision a scenario in which you could shed such loads in the interest of reducing peak energy demand.

The latest energy management systems can put those concerns to rest. It has become a truism that businesses will almost certainly will leave money on the table — not only now but in the future — if they ignore utility programs coming online that offer opportunities for energy savings. Food processors, for example, count energy as their third largest operating expense, and it’s expected to grow.

It’s helpful to address a few common myths, then walk through how an advanced energy management system (EMS) works.

Myths vs. Reality

Myth one: To save on my energy bill, I’ll have to sacrifice production.

Not if you take the right approach. With an advanced EMS, built-in rules and constraints protect the production process and always supersede energy savings goals. You define the curtailment limits for each load, with the assistance of an energy engineer, so you can be assured the load is protected. The priorities that you set determine how loads will be selected for curtailment.

Myth two: The system will complicate the existing controls and make them impossible to maintain.

Not if you choose the right vendor and technology. You can integrate energy management technology in a way that causes neither harm nor unexpected disruptions to existing systems. An advanced EMS for industrial businesses will have several features that ease integration.

For example, wireless connectivity will let you monitor and direct plant or facility loads without adding cumbersome wiring. For hardwired systems, the strategy is to control points as “normally closed contacts.” If there is a performance problem with the EMS, other systems will return to normal operation.

Embedded control setpoints and rules are tailored for your facility and for each load. Once fine-tuned, the system should run with virtually no operator attention. We’ve seen customers forget they were managing their energy and unwittingly disconnect the technology — it can be that noninvasive and low maintenance.

Myth three: If advantages claimed were real, everyone would be using energy management technology.

The advantages claimed are indeed real. The problem isn’t the technology; it’s the fact that there are disincentives to use it, plus a general resistance to change and lack of knowledge about current technology. A plant engineer won’t get fired over a high energy bill. But heads can roll if the frozen french fries don’t get out on time. So engineers who think an EMS is a risk to production won’t consider it.

The balance of concerns is shifting, though: With shrinking margins and tight budgets forcing companies to cut operational expenses, energy is increasingly under scrutiny as a controllable cost.

An advanced EMS makes possible four ways of controlling energy use and costs: demand control, demand response, real-time pricing and energy efficiency. Technical personnel often think it’s impossible to save energy and money while maintaining production. But demand control allows almost all facilities to use energy more efficiently by reducing usage peaks. To show how this works, we’ll consider as an example Powerit’s Spara technology, a hardware and software system that integrates with a facility’s existing controls.

Spara software resides on its own Web-based embedded Linux server. The server has a number of networking and communication ports for connections to various controllers (PLCs, microcontrollers, VFDs, and so on). As is typical with EMS setups, there is no additional hardware required. The software normally uses existing control points to manage loads. If there are loads in the system that are unautomated, the usual approach is to add them using industrial 900-MHz wireless units to minimize the additional wiring involved.

Continue on next page

Spara handles demand control by carefully choosing from a series of preselected loads. It then sheds the exact desired amount of power — or as much as possible given facility-specific rules. The algorithm it uses to make decisions about the frequency and duration of curtailment events contains two key elements: a predictive model that can be tuned to filter out intermittent demand spikes, and a calculation proven over time to select the best-fit loads needed to reach a prescribed curtailment level (so there are no unnecessary reductions in power).

In the case of peak demand, utilities calculate the kilowatts for each debit period by converting the kilowatt hours used into an average kilowatt demand. With a 15-minute debit period, the average kilowatt demand would be the kilowatt hours used in that 15 minutes multiplied by four. Spara monitors the actual kilowatt hours consumed by the customer’s facility as measured at the utility meter. It then automatically makes adjustments to meet savings goals (within constraints). Here’s how the process works:

1. The user enters a setpoint in kilowatt hours that the system will attempt to maintain for every debit period (usually 15 or 30 minutes).

2. During each debit period, Spara’s algorithm calculates the accumulated kilowatt hours for every subinterval (every 15 seconds) and the projected energy use for the rest of the debit period. If a potential peak is imminent, the system will reduce only the connected loads needed to keep the projected demand just under the setpoint. When the potential for a peak goes away, Spara releases the loads to normal operation.

3. Spara reduces loads in a priority order that the user determines. This can be set up as a static list, a logical list, a rotating list or a combination of the three. Many such loads typically are part of a production process. As such, each one is protected by constraints, or rules. These can be based on time, temperature, pressure or any other process variable. They can also be based on a combination of variables using logical dependencies that can be simple or complex.

A point to note is that the rules cannot be broken. The system will let the setpoint be exceeded if the protective rules require that loads be released. Simply put, the system will not shut down a critical production process in the interest of saving energy.

4. The Adaptive Demand Setpoint feature maximizes savings by ensuring that the system uses a setpoint that is neither too high nor too low. The user selects fairly aggressive starting points for each of 12 rolling monthly demand limits.

During the current month, the system will attempt to control demand to the preset level. Suppose the system exhausts the loads it is authorized to curtail and peak demand reaches a new high for the month. In this case the system will back the savings setpoint off by a preset percentage to a higher demand limit relative to the new peak. This is now the new setpoint for the rest of the month, which ensures that the system avoids unnecessarily controlling any power draw lower than the new maximum demand level.

A real-world example: foundry operations

Here’s an example of a typical demand control operation at a metal-casting foundry that’s controlling energy demand from furnaces:

1. Spara’s real-time algorithm predicts that the foundry’s current energy use will exceed its setpoint by 200 kW. The facility needs to shed loads.

2. The system determines which loads (furnaces, baghouse fans, etc.) are OK for reduction at this moment in time. These loads are available for curtailment.

3. Spara stages the curtailment actions based on the preferred order that’s been set in the system. Suppose there are furnaces A and B that have a priority of 1 and 2. Furnace A has 150 kW safely available for reduction, so Spara powers it down accordingly. It then powers Furnace B down 50 kW to get the remaining reduction needed.

4. Each furnace can operate at reduced power for only so long without disrupting operations, and that time has been set in the system. Spara monitors the reduction time and sees that Furnace A has hit that point. It releases Furnace A and further reduces Furnace B to get the rest of reduction needed.

Note that use of time as a constraint is a simple example of a rule that can be integrated into the system’s decision-making process. Rules can also be fairly complicated and logic based. (For example, if pump speed is X and tank level is Y then the agitator can be curtailed to speed Z.) In addition, they can be triggered by schedules or recipes (rules are different for strawberries and broccoli). The system will never break a rule to save energy.

Continue on next page

5. With the goal hit, all loads are released according the procedure set by the facility.

What happened here? The foundry’s processes were interrupted, but they weren’t disrupted. The changes were defined in advance as acceptable power reductions in return for energy savings.

All in all, an advanced energy management system not only enables demand control. It’s also key to participating in demand response programs, which pay you for reducing energy use by a specified amount when your utility or system operator demands it. That lets industrial users benefit (rather than only suffer) from the variability of real-time pricing.

Energy-intensive businesses shouldn’t view plugging into the smart grid as a to-do item for the distant future. It is increasingly essential for them to manage energy costs and consumption just to stay competitive.