Lessons from Successful SCADA and Automation Projects

Charles M. Burt1 and Xianshu Piao2

ABSTRACT

ITRC has designed, specified and assisted with installation, several dozen successful irrigation
district automation and SCADA (Supervisory Control and Data Acquisition) projects in
California, Oregon, Washington, Nevada, Arizona, and Colorado within the past 10 years. This
paper is intended to acquaint people who are interested in SCADA with key concepts, to reduce
their learning curve. The term “SCADA” encompasses many combinations and variations of
remote monitoring and control. An emphasis on good planning, with the use of high quality
equipment and expertise, will help guarantee a successful project.

INTRODUCTION

Irrigation Association members work predominately in the farm or landscape/turf aspects of

irrigation. Another huge arena of irrigation involves irrigation districts, which obtain, convey,
and distribute irrigation water to farmers. SCADA systems in irrigation projects have been in
existence for several decades. However, the vast majority of functional SCADA systems in
irrigation districts have been installed within the last 10 years. Many of the lessons learned are
applicable to on-farm and municipal SCADA system applications.

Why SCADA?

There are probably three major reasons why so many irrigation districts are investing in SCADA:

1. Irrigation must retire “art” and shift to an industrial control process, in which real-time
information is constantly used to make appropriate decisions. Reducing “art” in the
process fulfills the need and desire to:
a. Reduce diversions, to maintain in-stream flows in the rivers.
b. Provide more flexibility in water delivery to farmers.
c. Reduce pumping costs.
d. Conserve water and sell the conservable water.
e. Remove the mystery of operation details, so that new employees can be easily
trained, and so that managers can establish clear and measurable performance
guidelines for canal/pipeline operators.
2. There is often a need for automation that requires computers (Programmable Logic
Computers, or PLCs) at remote locations. Because it is the nature of computers,

1
Professor, BioResource and Agricultural Engr. Dept. and Chairman, Irrigation Training and
Research Center (ITRC). California Polytechnic State Univ. (Cal Poly), San Luis Obispo, CA
93407; [email protected]
2
Controls Engineer. Cal Poly ITRC. [email protected]

418
 electronics, sensors, and software programs to have occasional problems, it is prudent to
remotely monitor their performance at such sites.
3. Some districts have key trouble spots where water levels or flow rates historically get too
low or high. SCADA provides a means to remotely monitor those sites in real-time –
minimizing labor, vehicle mileage, dust, etc. while improving response time to problems.

SCADA CHARACTERISTICS

Some SCADA systems will be quite elaborate and involve automation, and others will simply be
able to transmit data from remote locations. However, all SCADA systems have the following
components, at a minimum:

1. A sensor
2. Some type of on-site apparatus that creates an electrical signal that can be transmitted
3. A local power supply to power the sensor and transmission unit
4. Some type of communication system, such as hard wire, radio, satellite, phone, etc.
5. A receiving unit on the other end of the communications
6. Some mechanism to display the information – which may be a simple alarm bell,
computer screen, message on a pager, etc.

The components listed above would provide “remote monitoring” – which is one-way
communication only. However, many systems (e.g., Figure 1) also include some type of control
capability – which requires two-way communication.

Communications

RTU/
Master PLC Actuator
Station

Sensor

Figure 1. Conceptual elements of an irrigation district SCADA system that involves control.

Many people use the term “SCADA” to denote the collection and transmission of data, plus an
automation process. An automation process may or may not require SCADA, so we prefer to
separate the two (See Table 1). Not all of ITRC’s SCADA projects involve automation, and not
all of our automation projects involve SCADA. That said, all of ITRC’s automation projects that
use programmable logic controllers (PLCs) also incorporate SCADA for remote monitoring,
alarms, and the ability to change target values.

419
 Table 1. Variations between and within SCADA systems
Frequency of Sensor Frequency of Data
Case Basic Function
Monitoring Transmission to Office
1 Alarm for high/low
Continuous Only if alarm condition exists.
Monitor values
As often as
Alarm for specific
2 once/second, as
values such as height, Only if alarm condition exists.
Monitor seldom as once/15
position, temperature
min.
Remote monitoring of
specific values such as As often as For river basins – often a few
3
height, position, once/second, and as times/day. For irrigation
Monitor
temperature. No seldom as once/day. districts – often once/minute.
alarming.
Once/day remote monitoring
4
Cases (2) + (3) 1/sec – 1/day can be over-ridden by an
Monitor
alarm exception at any time.
5 1/30 sec – 1/30 minutes.
Case (4) plus remote
Monitor Slower than 1/minute is
manual control of an 1/sec – 1/15 min
plus outdated and cumbersome for
actuator
manual operators.
Case (5) plus remote 1/sec – 1/15 min.
6 1/30 sec – 1/30 minutes.
changing of target With modern
Monitor Slower than 1/minute is
values for local, automation, 1/sec or
plus outdated and cumbersome for
independent more frequent is
Automation operators.
automation common.
1/sec – 1/15 min. 1/30 sec – 1/30 minutes.
7*
Case (6) plus feed- With modern Slower than 1/minute will
Monitor
forward between local automation, 1/sec or often not work with feed-
plus
controllers more frequent is forward. This is rarely found
Automation
common. in an irrigation district.
1/sec – 1/15 min.
8* Case (4) plus
With modern 1/sec – 1/15 min. This is
Monitor centralized
automation, 1/sec or rarely found in an irrigation
plus computation of
more frequent is district.
Automation gate/pump movements
common.
* - Denotes forms of automation with few examples of sustained success in irrigation districts.

It is common for one SCADA system to incorporate several of the cases in Table 1.

Overview of Building a SCADA System with Local Automation (Case 6)

An abbreviated flow chart for the process of building a SCADA system for Case 6 (of Table 1)
can be seen in Figure 2. Within each of the action blocks are numerous steps. Within the box
labeled “Perform Specific Field Tests”, for example, ITRC has several pages of procedures.
Likewise, for checking the wiring and calibration of the field PLCs alone we have 12 pages of
flow charts.

Components versus Systems

Figure 2 illustrates that successful implementation of a SCADA system, especially one involving
automation, is much more complicated than simply selecting a PLC and some sensors. SCADA

420
systems are composed of components that have hopefully been selected and connected to work
as a seamless system to satisfy specific objectives.

Selection of
Control/Monitoring Strategy

Surveying Data

Determine locations for control; Hardware selection

for gate or pump; sensitivity analysis;
development of performance specifications
Select communications mode(s).
- Local
- To office
- To operators in the field
Build Canal Model in Simulation Model - To others
Conduct
radio tests
Algorithm Selection
Decide upon
- Information needed at base station
Characterize the Canal Pools - Frequency of information updates
- Archiving
- HMI Screen displays
- Alarms
Tuning of the Control Algorithm in
- etc.
the Simulation Model

Specify PLC, sensors, SCADA

Characterize the Hardware for Controllability (cable

movement versus vertical gate movement; pump
speed vs. flow rate; etc.)

Develop PLC-language (Ladder, IsaGRAF,

etc.) program for insertion into PLC

Program the HMI. Install all

Develop Tag and Register Lists for
communications.
the HMI

Work with Customer and Integrator on HMI,

Use the SCADA initially only
PLC, Sensors, Sensor Calibration, etc.
in remote monitoring mode.

Perform Specific Field Tests and

Use SCADA for field de-bugging of
Verify the Adequacy of the Control.
automation.

Occasional troubleshooting and

modifications for several months after the
system is commissioned.

Figure 2. Outline for the Process of Design and Implementation of a Case 6 SCADA System.

The components must be carefully selected so that everything “matches”. For example,

1. A very good sensor might output a voltage signal, but that might be completely
incompatible with a requirement that the signal must be transmitted over a 3000’ cable.
2. A specific brand of water level sensor may be excellent, but if a 10 psi sensor is selected
when only a 1 psi sensor is needed, the resolution of the signal will only be 10% of what
is possible.

421
 3. 8-bit sensors connected to 16-bit computer input boards cannot give 16-bit resolution, or
visa-versa.
4. If a control algorithm requires an average of 60 readings/minute, a PLC that is only
capable of obtaining 60 readings in the last 10 seconds of the minute will provide
different control capabilities than one that can obtain one reading each second during the
60-second period.
5. The power provided to a PLC must be capable of powering all the sensors, radio(s), PLC,
heater, etc. in all weather conditions.
6. Sensors with proprietary software and communications won’t easily fold into a complete
system that can be updated and added to.

The examples show the following “matching” or “compatibility” requirements:

1. The hardware must be able to physically link together and communicate. This is the job
of the “integrator”.
2. The hardware must be compatible with the control/monitoring objectives. This is the
job of an irrigation automation control specialist, who must provide specifications to the
“integrator”. Control specialists from the chemical, electrical, transportation, etc.
industries have not been able to successfully apply their control logic to canal systems.

The emphasis with SCADA systems, in our opinion, is QUALITY, QUALITY, QUALITY – in
specifiers, integrators, software, and hardware. Even in systems with the best-quality
components, problems arise. The correct SCADA team can sort out those problems if there is a
willingness to work together and an understanding of the system in its entirety.

Many of the components and “systems” sold for the farm irrigation SCADA market are one-of-a-
kind, proprietary units of medium-low quality, that are not assembled or designed by a “team”
that looks at the expanded needs for a farm. They are typically designed to
measure/monitor/control one or two items that are sold by the same vendor that sells the
communication system and office software. They are often not compatible with other brands;
the result is that a farm may need many individual SCADA systems to accomplish what one
good commercial, industrial SCADA system could provide.

Other Consideration

Irrigation districts that have successfully implemented them typically quickly expand them and
wonder how they survived without them. But there are also many problem cases. Classic
problems are:

1. Cost overruns
2. Failure to achieve performance expectations
3. Failure to reduce operating costs in order to meet payback expectations
4. The thing just doesn’t work right

Secondary problems include:

422
 1. Scheduling errors
2. Interfacing problems
3. Incompatible equipment
4. Lack of acceptance
5. Adverse publicity

In our experience, the problems in irrigation districts have arisen because of one or all of the
following:

1. Entities are looking for a “silver bullet” that will cure problems quickly and with little
effort.
2. People focus on a few components rather than understanding that they need to consider a
system.
3. There is no clear plan for the present and future.
4. Irrigation districts decide to use the “local electrician” because he/she is a nice person and
dependable, instead of hiring an experienced integrator with ample successful experience
in irrigation district application.
5. Districts (or local government agencies) invent their own sensors, hardware, and
software.
6. Districts start too big, too fast.
7. Not everyone accepts the fact that problems will occur, and that there must be qualified
people to diagnose problems, service equipment and software, and stick with the
problems until they are solved. This takes an on-going budget.

As with any process, there are logical steps to follow in designing and implementing a SCADA
system. These include:

1. Master planning
2. Precise specifications
3. Vendor qualification
4. Vendor selection
5. Adequate training and documentation
6. User tools for future changes
7. Spares and warranties
8. Continuous and near-exhaustive testing
9. Realistic schedules

The process can take a long time before any actual installation begins. If the planning is done
properly, the installation can be accomplished in a few months. If the wrong people and
planning are involved, the installation may never be completed satisfactorily. A few of the
points above are explained in more detail, below.

Master Plan: A master plan identifies the need for automation, the degree of automation
required, what other features are desired, and the budget and cost justification. This represents
the guideline for all the work, so it is necessary to carefully understand which options are desired
and why. The plan must also consider the impact on operation, instrumentation, training,

423
installation, interfaces to other utilities (such as electric utilities), public relations, manpower
requirements, future expansion, and expected life of the new system.

Integrator: The component selection, matching, installation, and troubleshooting are so

complicated that most successful SCADA projects have utilized an “integrator” that assumes
responsibility for the complete package. The integrator generally understands communications,
sensors, human-machine interfaces (HMI), actuators, etc. and can make certain that everything
physically moves, measures properly, and communicates. It is extremely important to
understand that when SCADA is used in a canal automation scheme, the SCADA integrator will
rarely, if ever, understand canal hydraulics, simulation techniques, and control algorithms and
algorithm tuning. These are separate functions that require an additional expert.

HMI: The Human-Machine Interface (HMI) – the software and screens in the office - is
important. How easy is the system to operate? Are control and monitoring screens
straightforward and is information easily accessible? For example, an alarm condition that is
missed because it is mixed in with many nuisance alarms is as bad as one that is missed by the
instrumentation. Can in-house people make simple changes to the screen displays?

Reliability: Reliability is a measure of the system’s ability to minimize downtime by avoiding

failure, or at least to keep operating in a degraded mode by using special software and hardware
such as an uninterruptible power supply (UPS) and a redundant master computer. The tradeoff
includes weighing the extra cost of the additional equipment against the value of this function
and the likelihood of an outright failure. No machine will run forever and eventually you will
have to shut down your master at least for occasional checkups and preventive maintenance –
managers should be prepared to live with these tradeoffs or buy reliability up front. ITRC has
decided that for control variables in automated systems, it is essential to have a high level of
redundancy in sensors, power supplies, and A/D converters. Yes, this costs extra and may
reduce the number of sites that can be automated, but it ensures a better chance of success.

Maintenance: Maintainability is the ease with which fixes or changes can be made to your
system. In the case of hardware, consider what you can fix yourself, what spares you may need,
and how accessible the vendor is for factory returns and for minor upgrade contracts. Modularity
helps maintenance. Placing sensors so that they can be removed for cleaning/inspections, and be
replaced in exactly the same location without new calibration, is important.

SUMMARY

The potential exists for new and expanded SCADA systems in their many combinations and
variations of remote monitoring and control for irrigation districts and farms and large
commercial/industrial/golf irrigation systems. However, in order for customers to fully utilize
that potential, attention must be paid to all of the details – which, in many cases, can “make or
break” a system. An emphasis on good planning, with the use of high quality equipment and
expertise, will help guarantee a successful project.

424