DAF Coagulation Controller
DAF Dosing System

Pumps are almost ubiquitous in water processes; whether the pump is the main driver for water flow, is dosing chemicals, or is moving water around a sample line, there is (nearly) always a pump somewhere in the process. Often there are multiple pumps of varying sizes and specifications, making managing all of them a challenge for any operator.

Did you know that…
…today’s leading analysers incorporate the majority of control algorithms a water process should need?
…water quality controllers can often accept fault signals from pumps to help keep your water systems safe and efficient?
…Pi’s CRONOS^® and CRIUS^®4.0 analysers are the only analysers that can be linked using SMART and DIGITAL communications to a Grundfos DDA pump?

Analysers and controllers will often interact with pumps for a number of different reasons:

• Controlling chemical dosing – controlling a pump to dose a chemical into a system based on one or more measurements (e.g. chlorine residual and flow rate, please see the PID technical note for more information).
• Recirculation control – controlling a pump to manage the recirculation of a system (most commonly used in pools, please see the VSD technical note for more information).
• Failsafe protocols – complex water systems can go wrong for any number of reasons, and analysers can play a role in making those systems safer. Many pumps will output digital alarm signals when they detect a fault, alerting the dosing systems that something needs attention, and to stop dosing the chemical. Dosing into a water system which is not recirculating can lead to dangerous overdosing, and also wastes chemicals leading to increased costs.
• Duty rotation, or n+1¹ back-ups – in larger systems, pumps may be rotated in order to increase pump longevity or have a back-up system to rely on in the case of pump failure. Having a back-up, or n+1 system, can often reduce downtime significantly, and so is very useful in production lines where downtime can lead to large cost increases.

¹n+1 is a standard term used when considering redundancy. The ‘n’ indicates the number of items required to run the systems so, in this case, the number of pumps. The ‘+1’ indicates the number of back-up systems in place.

Here is a summary of the different kinds of signals that can be used to interact with pumps, some of the common applications they are used in, and their advantages and disadvantages.

Power switching relay based control

Description: Power On/Off provides power to pumps when the analyser needs the pump to be on, this can be 12-230V. Pumps are off and have no power when they are not pumping.
Advantages: Cheapest and most universal method of control.
Disadvantages: High pump and relay wear.
Commonly used for: Small dosing pumps, pools, small water systems.

Volt Free Contact (VFC) Control

Description: The pump is powered separately to the analyser, and only begins to pump when a low voltage control circuit is closed. VFC can be used with thresholds or with PID and can be used as on/off control or pulse width/frequency.

This means that the analyser can mimic a flow meter’s ‘Pulse Control’.
Advantages: Very commonly accepted by pumps and protects both the relay and the pump from high voltage surges.
Disadvantages: Not as precise as analogue control, especially for non-recirculating systems.
Commonly used for: Small to medium dosing pumps, pools.

Analogue Control

Description: Either a 4-20mA or 0-10V is scaled to match the pump output, and then is used to control the pump rate.
Advantages: Very precise control, even on non-recirculating systems. The pump is on all the time eliminating power surges and start-up fatigue.
Disadvantages: Can only be used on pumps that accept analogue inputs.
Commonly used for: Medium to large systems, large pool systems, any system where precise dosing is a necessity.

SMART and DIGITAL control

Description: Some pumps can now be controlled using digital communications such as Modbus or Profibus. These are designed to be used directly from PLCs and they give operators distinct advantages over more traditional methods. A digital signal has longer range, has a signal cable for 2-way communication, and is able to transmit far more detailed information than a simple on/off alarm signal.

An example of a pump with SMART and DIGITAL communication is the Grundfos DDA dosing pump with e-box add-on. The e-box provides a Modbus or Profibus connection for 2-way communication.

The CRIUS^®4.0 instrument controller from Pi is the only water quality analyser capable of utilising the full range of information available from the DDA pump. Information such as: total pump run time, pump pressures, volume pumped, and many more are now available to the analyser.

Error alerts can now be specific, telling operators exactly what needs to be done with their pump. This wealth of information can be useful in co-ordinating maintenance and increasing uptime for the pump and the system as a whole. All of this information can be stored, and data-logged, making the system very auditable.

Normally all of this information would only be available to operators with access to a PLC, and a PLC engineer able to integrate the two systems, but the CRIUS^®4.0 from Pi is a customisable controller that doesn’t need a PLC engineer to configure.

Control with the DDA e-box can also be extremely precise, with the pump more able to correct the pump output to very closely match the scale used by the analyser. A pump being controlled by a 4-20mA scale, asking for 60% of the pump’s output, was more than a litre off the real value of 60% of the pump’s scale. This means that on a 0-30l/h pump, a pump controlled by a 4-20mA signal calling for 60% of the pump’s output would actually be pumping 19l/h, whereas a Modbus controlled pump would be pumping 18l/h. This difference is small but measurable, and over a pump’s lifetime would add up to a remarkable discrepancy in chemical. It’s likely that with longer cable lengths, on a real site, this discrepancy would increase.
Advantages: The most precise control, excellent maintenance information, more uptime for processes.
Disadvantages: Currently only available with Grundfos DDA pumps and Pi analysers.
Commonly used for: Large to very large systems, or 24hr systems where savings on chemicals and reduced downtime are amplified. Food systems in particular benefit from increased uptime and excellent auditability.

There are many ways for analysers to interact with pumps, and no single way is a ‘one size fits all’ solution. Whether you are retrofitting to an existing system, or designing a new plant with the latest in SMART and DIGITAL communications, Pi’s instrument controllers have you covered.

Process Instruments supply CRIUS^®4.0 and CRONOS^® controllers with sensors for many single parameter systems like chlorine and pH but did you know that…
…Pi’s CRIUS^®4.0 and CRONOS^® controllers both allow the connection of multiple sensors?
…Pi’s multi-parameter systems can save you space onsite, are easy to use and have a number of communications options available?
…Pi’s CoagSense coagulation controller is a multi-parameter system that is now a key part of Pi’s product range?

An example of one of Pi’s most in demand multi-parameter analyser combines a CRIUS^®4.0 with a TurbSense^® turbidity meter, a pHSense pH meter, a ConductiSense conductivity meter and a UV254Sense UVA analyser with a temperature sensor on raw water systems. These have proven to be very popular in applications where companies and local authorities are looking to save on installation time, space and upfront costs.

Pi’s Multi-Parameter Controllers

Pi’s CRIUS^®4.0 Controller

The CRIUS^®4.0 is equipped with the capability to connect up to four sensors of any type with appropriate analogue outputs and relays. If four isn’t enough, don’t worry. The CRIUS^®4.0 can connect up to sixteen sensors by adding expansion boxes where needed, all using the same display and communications.

Equipped with data-logging as standard and multiple PID loops as options, the CRIUS^®4.0 is a cost effective alternative to multiple single sensor analysers, reducing cost while maintaining the highest quality.

Optional communications packages allow Profibus, Modbus ASCII, Modbus RTU, Modbus TCP, 4-20mA analogue outputs, and relays for alarms and control.

Customers requiring a no-frills controller should consider the CRONOS^® controller.

Pi’s CRONOS^® Controller

The CRONOS^® has the capability to control up to two sensors of any type with appropriate analogue outputs and relays. Equipped with optional PID control, the CRONOS^® is very able to control complex water treatment processes at a fraction of the cost of other controllers.

Advantages of Pi’s Multi-Parameter Systems

1. Space Saving

Finding space is becoming a real problem for installation engineers across many sites at the moment. As water supply companies are looking for more and more information and regulation increases, this has resulted in more and more instrumentation being installed. Most water treatment plants have limited wall space and finding room for new instruments is a real challenge. Both of Pi’s instrument controllers as multi-parameter analysers are solutions to this problem.

2. Easier to Use

With simple ‘plug & play’ inputs and outputs, intuitive display and button control, and with an individual manual configured to be the same as your analyser, the CRIUS^®4.0 and CRONOS^® make setup and ongoing use, simple and easy.

3. Communications

Modbus and Profibus communications are available which further cut down on wiring costs and PLC, SCADA input costs.

4. Remote Access

Pi’s remote access allows the user to connect their CRIUS^®4.0 through their local network (LAN) or a 3G/4G mobile network connection. This connection allows the user to have full control of their analyser from any computer, tablet or mobile phone.

CoagSense Coagulation Controller

The CoagSense coagulation controller combines a StreamerSense streaming current sensor, a UV254Sense UV analyser and a pHSense pH sensor to automatically control pH adjustment of the raw water and also the coagulant dosing in water treatment plants. A TurbSense^® turbidity meter can also be added if required.

The controller has revolutionised the industry where one controller can now monitor and automatically control your dosing pumps by sending a flow proportional signal direct to the dosing pumps, or to the plant’s SCADA. Previously, all of this had to be done by a PLC or a SCADA system which involved getting a specialist programmer on site and increased costs. Now there is a controller that does all of this in an affordable manner.

Process Instruments (Pi) are increasingly seeing water engineers and technicians relying on their PLC’s PID auto-tune feature. Many are becoming frustrated that the auto-tuners are unreliable when it comes to water processes. Sometimes the auto-tuned PID settings work, but then something changes in the process and the PID control no longer functions correctly.

In this Focus On, Pi aims to outline how PID auto-tuners work and why they are often unsuitable for complex water treatment processes. We also give a brief introduction to some of the principles of tuning a PID manually.

Did you know that…
…auto-tuners are not effective at tuning processes with long loop times?
…the CRONOS^® and CRIUS^®4.0 have many built-in PID safety features to make control easier and safer?
…PID loops can often be tuned remotely, meaning huge savings on site visits and travel time?

What is PID tuning?

A Proportional – Integral – Derivative (PID) controller is a control loop feedback mechanism widely used in industrial control systems. PID controllers allow a system to continuously and automatically modulate a control mechanism (e.g. a dosing pump, a valve, motor speed), to attempt to achieve a desired setpoint.

There are a myriad of ways that a PID controller can be ‘tuned’. In a very basic PID controller, the operator can choose how much of their control should be based on P, I and D. In reality, D is almost never required in water processes due to the nature of the control loop.

This is generally the task that ‘auto-tuners’ undertake. They make small changes to the ratio between P, I and D control, track the response, decide if the change is good or bad, and then change again.

What is the auto-tune doing?

So what the auto-tuner is doing is changing, measuring, and changing again; which can work perfectly on processes with a very short loop time (time between a change and a response from the system), the auto-tune making many changes in a short amount of time. The changes cumulate into a ‘tuned’ PID system.

On the right is an example system of a PID auto-tune.

In the diagram, we can see that for the system to work, the box in red is a crucial part of the process. If the change brought on by the PID auto-tuner isn’t visible quickly enough, it may affect the way the auto-tuner categorises the change (good or bad), or won’t give the auto-tune the opportunity to make enough iterative changes to bring a complex water system under control.

The good news

There is good news for water technicians and engineers and that is that Pi are here to help. All our salespeople are application specialists and are all capable of helping you manually tune a PID controller for your processes. Our CRONOS^® and CRIUS^®4.0 controllers are both capable of delivering excellent control in many processes, and have many built-in PID features to help make your process more controlled, robust and safe.

Where to start with PID tuning

In many ways, the auto-tuners mimic plant engineers making adjustments and tuning their PID settings. They simply lack 2 key components;

• Patience – because the loop times are so long, it will take time to tune a PID loop.
• Context – no auto-tuner can account for blocked dosing lines, or 2 identical pumps that are improperly calibrated and output different volumes, or seasonal changes in measured parameter concentrations. Context is incredibly important in tuning a PID loop.

With this in mind, here is how our engineers approach PID tuning on our instruments;

1. Check the measurement and output method are all working properly. This includes any calibration on the probes and on the pumps as well as checking that the dosing lines are clear.
2. Consider the loop time and change the PID’s Update Delay. This changes how often the PID algorithm makes a change to the output.
3. Start small and ramp up. Start with the P control and slowly increase the P factor until it is changing the measurement in a satisfactory timeframe. If you can get a good level of control with just P then that’s excellent. There is often no need to overcomplicate things past this. The addition of I and D components is often unnecessary.
4. If you have wild or erratic control lower the P until the control is stable, even if it doesn’t reach your setpoint then add some I control.
5. Always aim to have I as low as you can get away with, as I can cause more problems with control than it solves.

Once you have the basic P and I ratio sorted, or if you are unable to get the P and I into control, it is worth considering these additional settings which will help make your control scheme more robust and safer.

• Min/Max Output – restricts the range that the PID can work within, which can stop potential over- and under-dosing. This setting may also reduce the ability of the PID loop to respond to changes in the water.
• Start Mode – allows the process to be dosed using a flow proportional value or a manual percentage value, during a predetermined start up period. This is often used to get the process up and running before going into PID control.
• Ramping – smooths the start-up of the process where the error between the measured value and setpoint may be very large (which can cause erratic overshoot and overcorrection).
• Bumpless Transfer – smooths out the process of switching between manual and automatic control.
• Integral Wind Up Protection – puts a limit on the effect of I because I looks at error over time. A very small error over a large amount of time can result in a very large I output resulting in erratic control.
• Overfeed Protection – protects against the failure of other equipment, such as failing pumps or blocked dosing lines. This puts the control into alarm if the controller is calling for dosing but is seeing no change.

How our products can help

Even armed with all this knowledge, PID tuning can be a bit of a dark art. What works on one site may not work on another and even experienced engineers sometimes call us for help.

One of the ways we can help is using our remote access system Control InSite®. Our Remote Access portal (Control InSite®) allows people with the appropriate security clearance to make changes to settings and observe measurements from their desk.

Sales Manager, Dr. Rob Paramore, recently described his experience of changing a customer’s PID settings (at their request), whilst talking them through what he was doing on the phone:

Did you know that…
…PID can save you money by offering better process control?
…PID can help you maintain a setpoint, even with a variable process?
…the days of over complex and confusing PID are over?
…Pi can tailor a PID system to your exact requirements? You may never have to touch those settings again!

In this Focus On, Pi would like to introduce you to PID if you haven’t come across it, and discuss some of the useful advanced features of modern PID systems, like on Pi’s CRONOS^® and CRIUS^® models for those more familiar with PID.

What is PID?

PID is a mathematical tool created by engineers and is used in controllers. It is a feature often found in industrial controllers and is available in Pi’s controllers, as an inexpensive upgrade.

What is PID for?

The best way to explain what PID does, is to take an example. Most people have been to a swimming pool at some point in their lives, so this is the example we shall use. PID is also applicable in a huge variety of other processes. If you are not sure, you can always contact us to discuss your application.

When a person enters a swimming pool, they create a chlorine demand. They do this by introducing sweat, bacteria, organic molecules and other substances into the pool water. Chlorine reacts with these substances, which results in chlorine being used up and the chlorine level dropping. The chlorine level in this example, is often called the process variable or PV in the context of PID.

In order to maintain a concentration or level of chlorine, more chlorine needs to be dosed. If you dosed the same amount of chlorine per bather, the level would not be stable as all bathers create a different chlorine demand (e.g. swimming for fitness produces more sweat than swimming recreationally). Dosing manually brings in the issue of human error, and how operators approximate or calculate the amount of chlorine to dose based on current levels. Another issue with manual dosing is that it is not a continuous process, meaning it is unlikely a stable level will ever be reached.

What does PID do?

PID takes the measured level of chlorine or the PV and compares it to the desired level or set-point. This comparison gives the error which PID interprets and then calculates an output. The output is an electrical signal which controls the dosing of the appropriate chemical. The output can control heaters, dosing pumps and many other mechanisms that can be used to change the PV.

How is it used?

PID is made up of three parts, proportional, integral and derivative. Understanding what each part does helps operators choose the level of control best suited to them.

Proportional – Is the most commonly used for portion PID and suits most applications. When using proportional control, the further away the measured value is from the setpoint, the larger the output will be from the controller. This is an appropriate level of control for most processes, and users can gain a lot of control from a purely proportional system.

In some systems where the PV is lost to the process, e.g. chlorine from a pool, heat from a boiler etc., the proportional control never quite catches up with the setpoint. Users can see that although the process approaches the setpoint it rarely, if ever, gets to it. This is known as ‘droop’. The user can compensate for droop if the removal of the PV is fairly constant, simply by raising the setpoint, e.g. evaporation of chlorine from an empty pool. If droop changes often, (e.g. bather load or chlorine demand) then to eradicate the ‘droop’ then the integral part of PID can be applied to the signal to correct it.

Integral – The output from the integral term is determined by both the magnitude and the duration of the error. A small error over a long period of time will trigger a larger response than from a purely proportional system. This helps the elimination of the ‘droop’ seen in processes with continuous loss, and also serves to help reach the setpoint quicker.

Derivative – Derivative gain is rarely used and is generally set up only by expert engineers. Derivative gain uses the rate of change in the PV to try to predict future errors. This type of control is particularly susceptible to overcompensating, especially if there is even a small amount of signal noise (usually seen as spikes in the PV). Derivative gain is generally a tweak used by engineers to improve an already tight control, and is almost never used as an essential part of control.

What are the benefits of PID?

When properly set up, PID can lead to far tighter process control, which in turn can save you time and money. As an example, pool managers want to keep chlorine levels low, to improve the bathing experience and also save on chemicals. AquaSense is a chlorine analyser system that responds quickly and appropriately to a change in bather load (also known as chlorine demand). This means pool operators can save money whilst maintaining the safety of the pool. PID can also help reduce the risk of overshooting the desired setpoint, reducing the risk of dangerous overdosing of the chemicals.

Advanced Features and Safeguards

Whilst maintaining a setpoint with a PID loop is a huge advance over using threshold relays to maintain an upper and lower limit, it is sensible to control the loop with extra safeguards, such as:

• Maximum and minimum pump outputs. This is mainly used to prevent the controller from employing too aggressive a control, which can lead to overdosing. A minimum output can also be used in a system where the measured parameter is lost over time, to prevent the controller ever turning the dosing off.
• Ramp rate is a proportional control that allows users to choose how quickly or slowly the controller doses, in order to reach the setpoint. It is especially useful on startup, and can prevent the controller dosing too quickly.
• Wind up protection is an integral control, which limits the duration aspect of the control. This puts a limit on how much previous error can accumulate. Without wind up protection, there could be a very large integral value, if the process ever reaches zero or on startup.

These are all standard features in all Pi PID controllers.

Conclusion

In summary, PID is a very useful tool when used correctly, and can result in significant chemical savings, not to mention reduced pump wear and tear and lower electricity costs.