A case for a monitoring rethink in horticulture

"To make the point of difference quite clear, consider a standard nutrient delivery system", Geoff suggests. "Whilst there are lots of sensors, relays and servo motors, it is a closed loop system that is largely self-managing. Crucial to this is water delivery and often air flow. What happens if the power goes out? What if those pumps or fans fail subtly? Won’t those clever delivery algorithms try to adjust automatically? System failure is slow but inevitable as compensations are enforced by algorithms unable to rectify the real cause, which is not environmental but structural. This is very similar to computer-generated trades on the stock exchange. Regardless of fundamentals, automated sell or buy orders are triggered on settings that cannot compensate for the unusual. When it involves plants, the result can be catastrophic for no good reason." 

According to Geoff, the solution is quite simple: integrate complete infrastructure monitoring. "We like to divide sensors and controls into two categories: infrastructure and process. Most automation and control solutions do process very well, often involving highly scientific algorithms to optimise growth and management conditions. What they don’t do so often, if ever, is manage the infrastructure side of the equation." 

Therefore, such is the focus of this article, supplied by Geoff Schaller of Arcoflex USA. 

Before we look at how we might support remote monitoring, we first need to identify what components of our existing systems are at risk from under monitoring or even just lack of focus. Let’s look at the three most critical components:

• Mains supply – once the electricity goes, most hydroponic systems are in serious trouble. Knowing the moment power goes out can be critical to alert technicians or start up backup generators. This means that the monitoring system must have battery backup to support continued monitoring and alerting capabilities after the power goes out and some way to get the message out. Most systems we’ve seen have no such capability.

• Water supply – pumps sometimes catastrophically fail but more often, they degrade slowly. Filters block up gradually and sometimes, leaks emerge. Monitoring water pressure or flow alleviates this and prevents automation algorithms going into a spin attempting to compensate. It is a simple measure consistently overlooked in leading delivery systems.

• Air supply – many hydroponic systems rely on air flow through root systems to maintain humidity and moisture control. Similarly, in greenhouses, circulation fans are vital to distributing heat (or cooling) and maintaining drying patterns after sprinkling. Fan failure will almost certainly result in the emergence of fungal growth and inconsistence growth profiles. Protection is very simple with temperature differential monitoring and CT sensors monitoring current delivery to fans.

Another area of significance to hi-tech growing facilities is security. Firstly, you can’t have people just wandering around. Taps, valves and fittings are sensitive, and the careless knock could throw out those automation algorithms. Perhaps you don’t want people looking into your methodologies. So physical security, including even sounding alarms or forcing entry ways to lock, should be a major consideration with design automation solutions. The other key attribute of security is enforcing bio-security for pest control – winged creatures and pathogens on boots. Many compliance and accreditation programs for product sales across State and National borders require air locks and foot baths. So how you prove the airlock was always in the correct state? With door monitoring of course, yet how many existing plant monitoring solutions offer this? Very few. This is strange when the growth facility could be shut down simply because someone left the door open.

Another area that is a vital part of the supply that is often forgotten is the cool room. Your growth facility may be the best in the land, but if you do not store your harvested product correctly, you lose your accreditation or the right to sell produce. Spoilage or loss of shelf life is the outcome if your cool room is not pulling down well enough or not holding the correct temperature. Too cold and the product spoils. Too warm and the product degrades. Too much CO2 and the product blackens. So, you might spend $100k on hi-tech growing equipment, only to have the product rendered unusable for the sake of $5k spent monitoring your cool room.

The word remote implies that you can sit back at home or in the office and watch your equipment from afar. That, in turn, implies that you have internet connectivity to your farm. The curious thing we’ve seen is that most growing solutions do not provide an internet connection and rely on the end user to supply this. The problem with this is threefold:

• There is a cost and administrative complexity in setting this up and it isn’t easy. Finding data plans for SIMs is complex, especially if you need more than one. Most farmers and factories would rather the equipment supplier arrange this. Who wants multiple bills?

• Relying on a user supplied connection is fraught. If it fails, or the user fails to pay the bill, who is liable for resulting product risk? At which point does the risk owner change? How reliable is the user’s internet connection in the first place? When these systems fail, the first point of blame is the equipment vendor. Why not take that issue out of the equation?

• Is the owner’s internet connection backed up with an UPS? If not, you cannot even advise the owner that a power loss has occurred.

To our way of thinking, an appropriate remote monitoring solution must also provide the internet gateway, independent of the end user. Otherwise, the solution is sufficiently robust. Mostly, this is a 3G/4G solution as narrow band IoT solutions do not have the necessary bandwidth.

This is the big ‘C’ word. Control. You can have all the flashiest data in the world but if you cannot undertake basic aspects of control, you may as well ditch the hi-tech expense and live in your greenhouse. In some cases, the control aspect can be as simple as starting load balancing equipment. Or stopping it. Some simple examples might include:

• Start a backup fan when the main unit fails

• Flash a red alarm light (or klaxon) when a door is left open

• Activate heaters or coolers to control temperature

• Initiate watering sequences or prevent them

Note that we are talking about manual intervention here, not automation. Under automation, the algorithms involved are in control but what happens when those algorithms get out of control? Enter the human. He has a dashboard describing the circumstances and various buttons to press to initiate real world events in compensation and all from the comfort of his armchair at home.

"Given the non-scientific nature of infrastructure monitoring and control, it is easy and low in cost to implement. I think you have seen some relevant examples of what can and should be monitored, in parallel with technical monitoring. It seems that many solutions forget about the infrastructure around them, even though it is relatively easy to deal with", Geoff Schaller concludes. "It is time to adjust our thinking." 

For more information: