16/12/2024
This explainer to port automation kicks off a series of articles looking at how technology is changing the port sector, and how labour is lining up to resist these changes
So, why haven’t ports automated everything?
Some ports have. Increases in productivity are possible; but opinion is divided over whether the high cost of investment can pay back quickly enough to be justifiable in every case.
So it’s not “plug and play”?
Not at all. Port IT systems are bespoke, and no two operate in exactly the same way. Like any purpose-built industrial IT system, they take time to ‘bed-in’ and predicting everything which could go wrong at the design phase is virtually impossible.
This means that system errors and exceptions are addressed as they arise. While a good team can reduce the error rate over time by writing new code, a new update from one of the crane OEMs, for example, could break everything. Port staff will have to stay vigilant on a long-term basis.
Further increasing the complexity, an automated port must not only keep individual systems up and running, but those systems must be able to communicate with each other – the name given to this is Industrial Internet of Things (IIoT) – to ensure that vehicles are in the right place, and do not crash into each other. This adds another complicated layer to the development of a port automation system.
Which handling equipment can be automated at ports?
RTGs, straddle carriers, port trucks, and even ship-to-shore (STS) cranes – although the latter are the hardest to do, since there are differences in the geometry of the ships they are loading and unloading.
In some cases, automated equipment can present operational advantages not possible in human control. A straddle carrier, for instance, can be taller and can move faster, if nobody is required to be sitting in a cabin several stories off the ground. An automated port truck cannot endanger its fellows by being fatigued or hungover.
Why doesn’t automation increase port productivity in every case?
In Operational Technology (OT), algorithms decide what to do based on a sequence of conditions best understood by humans as a flowchart: if X is present, proceed to perform task Y. But when some parameter is off – for example, because is a container is not where it should be – this generates an exception, which can stop the algorithm in its tracks.
Unfortunately, in doing so, say if a vehicle is stood still when it is supposed to be moving, it generates an exception that other nearby vehicles cannot understand either, causing them to shut down.
This, incidentally, is why ransomware is so effective; it disrupts algorithms by generating exceptions, locking up equipment and preventing it from performing its functions. Computer programs – even so-called ‘AI-enabled’ ones – are incapable of improvisation, meaning that while a manual operator would be able to run their system in a partially diminished state, the OT system cannot.
How does this apply in real life?
So for example, experts have noted that automated ports are less capable in bad weather. A study by the US Government Accountability Office (GAO) found examples of terminals where productivity had actually reduced, since “…automated or semi-automated remote cranes… may not be able to operate in adverse weather conditions like rain or fog that obscure sensors’ or cameras’ views, whereas a human operator sitting in the crane may still be able to see.”
The report added that “…a few port stakeholders noted that automated equipment cannot respond to exceptions – unexpected deviations from its programmed operations – and may have more frequent maintenance issues, both of which require human intervention and decrease productivity.”
The same report, however, noted that automation led to greater productivity in most cases.
Why are ports always talking about 5G?
5G isn’t just for phones. In an automated port, machines like cranes and trucks need to be able to talk to one another via IIOT. A private 5G network within a port can improve connectivity over conventional Wi-Fi, which can be finicky.
On top of that, the distances involved are enormous compared with what your typical wi-fi network is designed to deal with – so a more robust connection via 5G can help with that, exchanging more data over further distances, and with fewer dropouts.
What is the ILA’s problem with automation?
Trade unions exist to constantly agitate for better pay and working conditions for their membership, and protect them from job cuts, exploitation, outsourcing and offshoring. A labour union which for whatever reason cannot, or will not, do this is an ineffective one.
With the advent of containerisation, ports went from boxes and bags loaded, stowed and lashed by hand, to containers hoisted by crane, causing many job losses. Automated port trucks, cranes and gantries likewise propose a threat to the jobs of ILA stevedores.
Could ports not be part-automated, part-human-operated?
This is the case for automated ports. Humans monitor the process from a control centre. One human may be in charge of continually overseeing several pieces of equipment, or a team of humans-in-the-loop, as they are known, may troubleshoot individual ‘exceptions’ with several pieces of equipment as and when they arise.
But when it comes to the workers on the ground, it is an all-or-nothing matter. It is not safe to put human workers in a place where equipment could start moving without warning – hence why automated areas are segmented off from human foot traffic. But this also works the other way around. A system with 10 human-operated straddle carriers and 10 automated units would be extremely difficult to programme, as the erratic or unexpected manoeuvrers of the human drivers would generate so many ‘exceptions’ that the automated vehicles quickly become inoperable. (This is the main problem with self-driving cars, as well).
But port operators say that automation does not necessarily reduce jobs. How can this be?
This is based on some selective reasoning. Automation could increase port throughput to the extent that more jobs would be needed to accommodate this. But as we have seen, port automation does not bring the expected productivity gains in every case. For example, the aforementioned GAO report notes that while automated ports can operate round the clock, in some cases, “…extended hours of operations would not improve performance because trucks do not want to come to the port to pick up containers at all hours, or because warehouses are not always open for trucks to drop off the containers”.
And clearly, ports cannot grow ad infinitum: their growth will be linked with, and limited by, shifts in the global economy. Nor is there necessarily a link between trade growth attributed to more efficient ports, and the prosperity of ordinary Americans – with the gap between wage growth and US economic productivity having opened up in the 1970s and grown wider ever since. Actual salaries have increased by around 30% since 1978, a change attributable almost entirely to the boardroom.
Why has Trump sided with the unions? Isn’t he a Republican?
Mr Trump has “studied automation”. He claims to know “just about everything there is to know about it”, which is certainly more than your correspondent could boast. He equates US port automation with malign foreign influence, saying that “…foreign companies have made a fortune in the US by [gaining] access to our markets”.
“They’ve got record profits and I’d rather these foreign companies spend it on the great men and women on our docks than machinery, which is expensive and which will constantly have to be replaced.”
Port machinery tends to be designed for a lifespan of around 30 years or more whether or not it is automated. But Mr Trump is not the only US politician to voice concerns over the provenance of American port equipment, with Chinese crane manufacturer ZPMC coming in for accusations of spying recently.
