To the Right Level

As Peter Welander explains, analyzing your application is the first step in selecting the right level sensing technology

Measuring the amount of a substance, either liquid or solid, in a container is one of the most basic elements of process instrumentation and goes back to time immemorial. While we generally talk about measuring the level in a tank, in most situations the actual information desired is volume. Situations where a level measurement is desired for itself are far less common. As with flow, level measuring techniques are highly varied and use many technologies. For the purposes of this article, we examine only those techniques that are adaptable to electronic data collection and where multiple manufacturers are commonly available.

Choosing a level sensor starts with defining the specific process need within the application constraints. Level measurements are either “continuous” or “point level: (discrete). That is, they either indicate whatever level is in the tank, or whether it is above or below a specific point.

If you need 100 liters of a liquid for a process from a given tank, it may be enough to know that the measure is greater than 100 liters, but how much greater isn’t a concern. Similarly, you may want an alarm if a level has fallen below a specific point before a tank is pumped dry, or a high level alarm before it overflows.

Bear in mind that continuous methods can be used to perform point level functions, and multiple point level measurements can effectively provide useful, if somewhat crude, continuous data.

Begin with analysis

Selecting a measurement approach should begin with an analysis of the process and what information you need:

• What are the contents?

Measuring a solid has specific challenges since powders and granular products don’t always settle. They can coat interior surfaces or cause “rat holes” near exit points. These can make for unreliable readings depending on the technology. Well-behaved liquids are much easier to deal with, however, slurries and solids content can also cause settling issues in liquids. Moreover, foam, turbulence, or even dust can deceive reflective measurements, and dielectric characteristics can affect capacitive sensors.

• How precise a measurement?

This usually only affects continuous measurements, but depending on the size of a vessel, very precise measurements are possible (±<1%) but these are likely to be expensive. The need for precise readings from a large tank over a wide range is rare.

• Is contact with the contents possible?

Many methods involve being inside the vessel and in contact with the contents. Others do not require this.

• Is it possible to penetrate the tank wall?

Some measurement methods do not need penetration into the tank at all, and so are appropriate for when the contents are very sensitive or if the tank is under high pressure. If the situation calls for adding measuring capability to an existing tank and there is no appropriate spud or port, modifications may be needed.

• Do you know the interior dimensions?

If volume is the ultimate goal, all interior dimensions must be fully quantified, with deductions made for baffles, paddles, heat exchanger elements, etc. A precise level reading will not correct for imprecise dimensions.

The biggest dividing element for choosing level sensing technology is the tank contents. For purposes of this discussion, we define liquids as a substance that will settle to a uniform level and flow through a pipe. A solid might flow to some extent, but will not necessarily create a uniform surface and may not drain off walls. A liquid that is very viscous or has heavy solids content may behave more like a solid.

One of the simplest and most reliable ways to determine the contents of a tank is to weigh it. This is the only method that gives an actual mass reading and it is not dependent on knowing the interior dimensions of a tank. Load cells placed under the feet of a tank can determine its contents after deducting empty weight. This works for any type of contents and, if there is no interference from piping or other connections, the method can be very precise. There are practical limits to overall size, but don’t overlook this obvious approach.

If this is not a practical solution, all others become more complicated and suitability depends on a combination of requirements. A summary of the various level technologies is provided below.

Electromechanical solutions

These approaches have some type of moving part as a common element, either a float that rests on the surface or device that has to move through the contents.

Using a float on the surface of a liquid is a simple and reliable method of measuring its level, as long as there is nothing in the contents that could interfere with free movement. There are many valves, switches, and encoders activated by floats which can provide point level or continuous readings over a narrow range.

Some of the most sophisticated continuous measuring float designs use magnetostrictive sensing technology. The float is doughnut shaped and rides on the outside of a tubular waveguide. The waveguides can be quite long, (up to 15 m), so they operate in very large tanks. The float contains a permanent magnet which causes a disruption in an electrical pulse sent down the waveguide. The equipment can measure the point of the disruption with remarkable precision and repeatability, often ±<0.02 mm. Once units are installed and set, there is no additional calibration required.

“Magnetostrictive technology is excellent for wireless communication,” says Mike Geis of Ametek Automation & Process Technologies. “Response to the pulse is immediate so current consumption is minimal.”

The waveguide can support two floats, which makes magnetostrictive measurement one of the few technologies capable of continuous measurements of layers of liquid (e.g., oil on top of water) with one device. “As long as there is at least a 0.1 difference in specific gravity, we can design custom floats to measure multiple liquids,” Geis adds.

Vibration and paddle wheel are two point level methods which are similar in that they insert a probe that moves into the contents. A vibrating probe inserts something like a tuning fork into the material that is set into continuous vibration by a piezo-electric crystal. If the probe is not buried in the contents, it can vibrate freely. If it is immersed, it will not vibrate properly, which the mechanism can recognize and signal accordingly.

Similarly, a paddlewheel uses a motorized paddle blade or flag shaped fin on a shaft connected to a small motor. If immersed in solid product, the device can’t turn and sends a signal. When the contents are drained, it can rotate. Both these methods are intrusive and subject to damage from contents.

Like mass measurements (weighing the tank), pressure and differential pressure methods determine level by measuring pressure head at the tank bottom (or wherever the gage is located). If a tank is vented to atmosphere, a simple pressure gage is sufficient. However, if the tank is closed and can be pressurized or de-pressurized, a differential pressure reading between the bottom and open space in the top will self compensate for any difference between internal and atmospheric pressure. This method works well, but it adds plumbing.

Electrical probes

When a point level measurement is sufficient and product contact is permitted, capacitance and conductivity probes provide uncomplicated and reliable solutions.

Conductivity probes are simple and provide point level readings with conductive liquids. They are often mounted in clusters of two or more to measure high and low levels. If a liquid is non-conductive, another approach must be used.

Capacitance probes determine the presence of solid or liquid content due to a change in capacitance around a probe with multiple electrodes. A radio frequency current is fed into the electrodes, and the change can be measured based on the dielectric value of the substance in contact. Some designs can determine the dielectric value of the product, which means they can differentiate between different contents. For example, a probe immersed in oil will give a different reading than if immersed in water. This can help when dealing with situations where there is more than one product possible in a tank.

Some capacitance sensors can read through the wall of a non-metallic tank, allowing it to take a point level reading without a penetration or product contact. The sensor can mount flat against the tank wall, or wrap around non-metallic pipe. If the tank is metal, the sensor can be placed on a sight glass or tank well made from plastic pipe.

“Capacitive sensors work great on a lot of solids,” says Roger Saba, Product Manager for Turck Instrumentation Group. “There are some odd materials, such as chlorinated detergents, that can coat the inside tank wall and change the characteristics of the plastic. Some capacitive sensors will lock on and give a false reading, but more sophisticated units can see through that.”

Thermal probes insert into the tank. They use a small heating element to warm the tip and measure the temperature increase. If there is no liquid surrounding the probe, the temperature increase will be relatively large. However, if there is liquid, the heat will dissipate and the probe will not warm up as much.

Light beam sensors look for a blockage between a light transmitter and receiver. The presence of a solid or liquid blocks or dissipates the beam, indicating its presence. This technology depends on the ability of the product to drain and not block the light, indicating false positives.

Electrical reflective

Ultrasonic and radar techniques are useful in that they work in a wide variety of applications and do not require product contact. While they do need an access point to the tank (except for through-wall point measurement), it is at the top. Both technologies have capabilities to see through different kinds of internal obstructions, and each has its own limitations.

Ultrasonic technology can be used in multiple ways, which makes it very versatile. A sound pulse is sent into the tank and the sensor examines time for an echo to return. Taking humidity and temperature into consideration, it can calculate distance to the surface. Ultrasonic measurement has some problems with dust and foam, but this varies by application. It also has a relatively limited temperature and pressure range compared to radar.

A different type of ultrasonic sensor can be mounted on the vessel wall and take a point level reading without penetration. The sound pulse and echo can determine if there is solid or liquid material on the other side of the wall. In some cases it can differentiate between filling beyond its level and merely a coating of sticky material. This approach is particularly useful where capacitance measurements are not possible and there can be no product contact or even penetration of the tank.

Radar technology has been on the scene for more than 25 years, but has grown in popularity recently as its capabilities improve and costs decline. In the past, their steep price, large size, and high power consumption put radar sensors only in the most critical applications, but now they are used in an increasing variety of situations. Radar is similar to ultrasonic, but has fewer restrictions and is generally more precise: a microwave pulse has better capabilities to see through foam and dust, and fewer problems with pressure and temperature.

Radar sensors can be configured for non-contact applications, or with a waveguide that extends into the contents. The non-contact design is more common, but a guided wave configuration helps in situations where liquids have a very low dielectric constant and do not reflect the microwave signal well. “Guided wave designs have a probe to get the energy down to the liquid and back,” says Boyce Carsella, Radar Product Manager for Magnetrol. “The return signal strength is much higher if product contact is not a problem.”

Radar sensors are especially useful in reactors where there are high internal pressures, high temperatures, mists, vapors, turbulence, and other problematic conditions. “The biggest issue is foam,” says Carsella. “The questions we ask are, what is the dielectric constant of the liquid, what is the bubble size or density, and how thick is the foam layer?” He advises that guided wave designs are better able to see through a heavy foam layer.

The wide variety of antenna configurations for radar sensors allows you to choose the best approach for the interior space and liquid characteristics. Moreover, various frequencies offer specific capabilities for difficult situations and liquid properties.

Nuclear (radiation) sensors

While this is a very effective solution, its expense and specialized requirements usually makes nuclear sensors a technology of last resort. The approach is very simple: a radioactive source that emits gamma rays is placed on one side of the tank. Sensors, similar to Geiger counters, are mounted on the opposite side to take point level readings. The tank contents, solid or liquid, absorb gamma rays in a predictable manner that allows the electronics to determine the level. The measurement is only as precise as the number of sensors, so it is typically used for high and low limits.

Radiation sensors do not require any product content, nor even a tank penetration, so this approach is valuable for high pressure/temperature applications, valuable products, and where it is impractical to modify existing equipment. However, having a radioactive source capable of penetrating a typical stainless steel tank requires specialized types of licensing and training for operators, so this approach requires careful forethought.

Ongoing developments

Like most types of process instrumentation, level sensor vendors are looking for more efficient ways to use power and reduce electrical consumption. Some approaches are better than others in this respect.

Low power consumption generally makes sensors more practical to interface with wireless transmitters and improves capabilities in hazardous areas. Most can be triggered intermittently, but the ability to respond quickly and not use a lot of power in the process varies a great deal.

Sensors using magnetostrictive technology for example, have greater potential for such adaptation than paddle wheel or thermal approaches. Of course electrical consumption is only one piece of the puzzle. All other performance trade-off considerations still apply when examining an application.

---

Radar Under Pressure

The speed of radar signals can be affected when confronted with high temperature and pressure conditions, resulting in significant process measurement errors that need to be rectified, says Rob Vermeulen

Radar technology in general has been introduced to the process industry as being a measurement technology using high frequency electromagnetic waves that are not influenced by the gas phase it travels through and the temperature and pressure conditions in process vessels.

But as processes get more extreme in temperature and pressure it is time to have a closer look at radar behavior in those critical applications and the solutions available on the market that overcome the obstacles.

Speed of light

All radar technologies available on the market that are used to measure level use the “time-of-flight” principle. This means that the radar measurement device measures the elapsed time between emitting and receiving of a pulse consisting of a bundle of high frequency electromagnetic waves. The frequency of the waves vary between 1 GHz for guided wave devices, and 6-26 GHz for free space radars.

Radar signals travel at the speed of light when traveling through a vacuum. This speed, however, can be affected when not traveling through a vacuum, with the pressure and temperature of a specific gas phase or liquid having influence.

The amount of influence depends on how polarized those gases are - in other words, how much the dielectric constant changes. Hydrocarbon vapors show little effect even under high temperature and or high pressure process conditions. But a high polar steam does. The Dc (Dielectric constant) of steam at 212°F is 1.005806. But at 691°F it is already 3.086.

Steam slows

In a typical steam application, the level of the water in a condenser or boiler is of utmost importance. Radar measurement devices are used more and more in these critical applications. They offer great benefits – advanced diagnostics, and insensitivity to build-up and temperature fluctuations.

But steam is a highly polar gas, which means that the speed of radar signals in high pressure and temperature applications are subject to a reduction in speed. In a boiler for instance, this leads to a lower water level reading then there actually is.

This can be dangerous and influences the performance of boilers and causes a reduction in the quality of steam. The error can easily be as large as 30-40 percent depending on the pressure and temperature of the steam and distance from the launch of the signal to the actual water level.

Overcoming the effects

The simplest (but not the best) way to overcome the radar speed slowdown problem is to put a fixed offset in the measurement device, by simply inputting the temperature or pressure and having the radar unit calculate the “offset”.

The problem with doing that is there will be rather large “errors” during the start-up of an installation. The normal operating conditions have not yet been met and thus the unit will be over compensating. One could also program a compensation table in a DCS or PLC and connect this to a pressure or temperature transmitter.

The most accurate method, however, is through a “dynamic compensation” circuit on a guided wave radar. Here, a reference signal at a known distance is used to compensate for the delay in speed of the radar signal measuring the water level.

This is done dynamically, which means that when, for example, the reference pulse signal shows a small shift in time, the level signal will be compensated for this small shift. In converse, if the reference signal shows a large shift, then the level signal will be compensated for this large shift.

Thus, to summarize, the use of radar signals for level measurement in high temperature and high pressure applications and especially where steam is involved is not always straightforward. Under these conditions the speed of radar signals can change causing large measuring errors. One solution is the Endress+Hauser Levelflex guided wave radar, which is able to effectively compensate for changing radar signal speeds and so increase confidence in the accuracy of level process measurement.

Rob Vermeulen is Product Manager, Level Instrumentation, Endress + Hauser Inc.

---

Typical level sensing methods

Radar level instruments

Levelflex guided wave radar