Eliminating the hype and snake oil, there’s a lot to see for manufacturing in IoT. And believe me 99% of this totally unexplored territory. I am writing this huge series of blog posts on IoT, using the opportunities that I see most stakeholders let pass by as they aren’t even aware of the existence.
IoT in manufacturing = No revenue leakage!
3 years back, I visited this manufacturing facility in Illinois, where I met 13 key stakeholders to learn about their concerns surrounding how they carried about manufacturing operations and how they actually saw IoT help them overcome these challenges.
After 4 days of meetings and interviews, what I realised was the fact that there was a lack of understanding in terms of their own machinery.
I suggested a plan for discovery using hardware product development techniques that involved starting with hardware prototypes to start tracking and observing for potential “Revenue leakage”. If you’ve ever work with me, you would hear this word a lot of time from me. Revenue leakage is one of the leading ROI metrics for IoT implementations.
Everything on the shop floor started talking to us within a month and our demo software systems that were connected to these hardware prototype started reflecting just how chaotic a plant floor can become. When we walked the stakeholders through this, they were shocked to their core to see what they could’ve never imagined.
Despite of an award winning staff and a production planning team that was second to none, they had a 37% unexplained downtime.
What happened there was pretty complex, but here are some parameters that we were able to measure and track to identify revenue leakage in this facility:
- Manufacturing cycle time
- Overall equipment effectiveness
- Production yield rates
- Perfect order performance
- Return order authorization
Now, all of this wasn’t visible when centralized decisions were the key to effective production strategy. This is how fascinating Industrial IoT really is!
Let’s now take look a into IoT implementation at a much granular level in Manufacturing. I hope that you can easily relate to these use cases. If you can’t, hit me up with an email and I will be more than happy to share what you can do with IoT in your industry.
Gas Leakage Detection using Ultrasonic sensors
Acoustic monitoring services use ultrasonic sensors to detect leaks based on changes in the background noise pattern. These sensors respond to the sound generated by escaping gas at ultrasonic frequencies. The ultrasonic sound level is directly proportional to the mass flow rate (leak rate) at a given distance. The leak rate in turn is mainly dependent on the size of the leak and the gas pressure. Most gas leaks – as well as operating mechanical equipment and electrical emissions – produce a broad range of sound that span from the audible to the ultrasonic range (approx. 20 Hz – 10 MHz). The ultrasonic range itself ranges from 25 kHz to 10 MHz.
Gas leakage detection usually has ranges like these when it comes to ultrasonic sensing:
Minor leak : Greater than 0.1 kg/s with duration less than 1 min
Significant leak: Between 0.1 and 1.0 kg/s – duration with between 1 min and 5 min
Major leak: Less than 1.0 kg/s with duration over 5 min
The range in which these Ultrasonic gas sensors can detect leakage is in a radius of 8-12 m. The lower estimate of leaks detection for these sensors is in range of 0.03kg/s if the detection coverage is reduced to 4-8m.
You can make this using:
Low fidielity Prototyping solution:
XzArduino/Bosch XDK + Zigbee + MQ5/HC-05
A better solution:
Bosch XDK + Zigbee + E4C-UDA
Now, if you are wondering on why I didn’t suggest BLE here, it is because it isn’t optimized for flammable gas detection. You can read more about it here in Wireless IoT protocols for hazardous and flammable environments.
Self-Powered Wireless Vibration-Sensing System for Machining Monitoring
Wireless sensing systems are used for sensing, analysing and monitoring the temperature of Milling cutter. This prevents machine failure due to overheating and thus increasing the reliability of the tool.
What about the maintenance due to battery replacement?
While researching and experimenting I found that maintenance becomes difficult due to the issue of battery replacement when many sensors are used in the network.
Solar cells, solar power came to my mind when we think of energy harvesters. They are doing wonders in the real world.
But solar cells are not preferred for indoor applications especially for monitoring the machine in any factory.
Bad idea – Piezoelectric energy harvesters, because they have more complex circuit.
Good idea – Electromagnetic energy harvesters.
Look at the below diagram which shows the conceptual approach of how to monitor a milling process using attachable energy harvester powered wireless system.
Lemme explain this in detail..
The self powered system consists of :
- Electromagnetic energy harvester
- MEMS accelerometer
- Wireless module – Zigbee
Energy harvester consists of magnet and Inductor. Inductor is attached on machine and several permanent magnets are attached on the spindle of the machine.
During the milling process the Inductor is fixed on machine while the magnets rotate with spindle. This results in continuous relative motion between fixed Inductor and rotating magnets.
This simply means that a periodic alternating power output is generated i.e. mechanical energy into electrical energy. In return the harvested electrical energy powers MEMS accelerometer and Zigbee module.
And a self powered vibration sensing system is formed and is able to sense the vibration and transmit it to the terminal computer.
Vibration of Induction Machine due to the rotor imbalance
The figure below shows the wireless health monitoring of an induction motor.
The vibration signal from three accelerometers is recorded and stored at a base station. These signal can further be used to extract detailed information of the health of induction motor.
Let us understand this in detail –
The main circuitry would include power supply, ADXL330 MEMS Accelerometer and CC2430/31 Zigbee module
ADXL330 MEMS Accelerometer is 3-axis acceleration measurement system built on a single monolithic IC. This accelerometer consists of a micro-machined sensor and signal conditioning sensor and has a measurement of +/- 3g.
And CC2430/31 Zigbee module combines the performance of CC2430 transceiver and 8051 microcontroller with 128 KB flash memory and 8KB RAM.
ADXL330 MEMS accelerometer (mounted on the motor housing) + CC2430/31 (8051 microcontroller + Zigbee)
Wireless Monitoring of Winding Roll Pressures
To measure near-core roll pressure during the winding process, I inserted thin film pressure sensors in the layers near the core and tethered to a wireless data transmitter to send the roll pressure data to a nearby computer over Wifi frequencies.
A system for Inverter protection and real-time monitoring
All of us know that it is very important to protect the inverters. And since long many ways have been employed to protect the inverter with special protection devices and control circuits, fusing being one of them.
So what is the need to replace them?
Because fuse has relatively slow response time which means that additional protective equipment like crowbar circuits would be required. While working on it I also tried suppressing the load-side transients and DC supply with filters, but this increases cost, weight and power losses.
So, I proposed an inverter protection and monitoring system.
I replaced the shunt resistors by Hall-effect-based sensors as they have isolation from main power circuit and independent of the parameters like dust, humidity and time. They have an added advantage of wide frequency bandwidth, low temperature variation, therefore they are ideal for current detection.
An overcurrent protection circuit for every set of parallel connected MOSFETs.
Intel 80C196KC (MCU) + Zigbee / Bluetooth
Detection of Overheated Rollers in Belt Conveyor Systems
Multiple production lines stop due to an unplanned downtime. Fire at belt conveyors are very common which is caused by bearing failures. Due to the bearing failure rollers heat up and ignite when belt at the conveyance stops.
There are available roller condition monitoring systems but they are quite expensive. So here I show you an economical yet effective method I used for detecting overheated roller attached to the conveyor belt.
Solution is here for you.
A single sensor which consists of multiple discrete sensors allows to monitor all rollers in a belt conveyor system. The picture below shows a data transmission unit and discrete temperature sensors of the sensor system which move over all the rollers of a conveyor belt system.
Batteries and electronic components are located in the belt’s rim because of lack of space.
And data transmission takes place through Bluetooth Low Energy which automatically connects to one or more base stations.
And the sensor system?
The sensor system consists of thermopile because through thermopile we can measure very small temperature changes and has a resolution of 4,9 µK. This simply means that the step size of 4µK can be achieved.
Online Monitoring of Shuttle System and Bearing Condition
Conveyors or rails are used as shuttles to move the load in industries which move up and down to lift the heavy load like engines, chassis etc. Number of shuttles are connected to a single production line and the successful alignment is in horizontal parallelized. The load is not distributed equally when the shuttle is in unparalleled condition. Thus affecting total production line.
I found out that due to lubrication failure, bearing is affected by high temperatures and with continuous usage the bearing becomes hotter and eventually gets damaged.
To monitor bearing temperature, I used heat temperature sensors and an alarm rings whenever the temperature overtakes the withstanding temperature of the bearing. Displacement sensor is used to measure the displacement which is caused by continuous movement.
LM35 is used as a temperature sensor and fixed it near the bearing. It has more memory, processing and communication capabilities.
Ultrasonic sensor and displacement sensor measures the displacement of the shuttle.
ZIGBEE Based Parameter Monitoring and Controlling System for Three Phase Induction Motor
It is very important to protect the induction machines to ensure smooth functioning of equipments attached to them. While doing so I faced a challenge – cost of installation. And yes, it’s very impractical to manually monitor these machines.
Overcoming these limitations we have come up with a cost effective and a simple system which is based on Zigbee.
The below diagram shows the transmitter system and receiver system.
Transmitter system consists of sensors, transducers and microcontroller to acquire voltage, current, temperature and speed of induction motor. These values are sent via microcontroller and Zigbee- wireless connection over a system where they are compared with set values. Controlling signal is sent in case the measured value exceeds the set value. This would result in adjusting the speed, turning the fan on or stopping the motor.
And the receiver module simply consists of the Zigbee module and the monitor over which the output is compared.
LM35 (temperature sensor) + AT89C51 (speed and vibration sensor) + ATmega16 + Zigbee
Detecting Nanoscale Vibrations in Rotating Devices
Inadequate dynamic behaviour of a positioning system affects the dimensional accuracy of manufactured parts. The appearance of vibrations can cause unwanted motion in any axis. The dynamic forces that arise during the rotation of devices, such as the spindle of an air bearing, reflect these unwanted movements.
352C15 from PCB Piezotronics, which has a sensitivity of 10 mV/g and a bandwidth of 12 kHz
Coolant Ageing Monitoring (Grinding and Polishing Machinery)
Changing coolant characteristics can lead to the development of microbial or bacterial bio-hazardous load during metal grinding process. This gets unsuitable for use and ageing also causes some health safety issues which are associated with skin irritation and respiratory problems. Therefore to meet the end manufacturing specifications and comply with the health and environmental regulations it is very important to monitor the coolants.
Coolants are alkaline in nature so that they can neutralise the acidic nature of bacteria or any other biological content which starts to grow in the fluid tank. This simply means that the pH keeps on changing.
I studied 2 coolants- Multan 61-3 DF by Henkel (red line) and Syntilo 81BF by Castrol (black line). Below is the graph which shows the change in their pH.
Output of the photodetector can be directly digitized and processed in the microcontroller unit (MCU) of the wireless node, without the need of additional interface circuitry.
The wireless node is hosted in an IP54 environmentally protected enclosure and transmits the sensors data through a ZigBee network in the 2.4 GHz by using a ZigBee to WiFi Gateway wireless communication unit.
Polymer fiber sensor + Zigbee
Traction Motor Condition Monitoring
Mechanical imbalances majorly result in the mechanical failure of motors. Rollers and Bearings are affected by a continuous stress due to lack of lubrication and installation, and corrosion.
This unduly faults can be sensed through vibration sensors and temperature sensors as faults result in increase in vibration and unbalance shaft current. Temperature can also be increased by any type of bearing failure.
ADXL345 three axis accelerometer measures vibration, DS18B20 measures temperature and ACS712 measures current. This information is transferred via ESP8266 IoT Wi-Fi chip which is powered by a small Lithium ion battery.
Fugitive Emissions Detection
The emission of vapours and gases occur from pressurised equipment which may occur due to leaks or other irregular release of gases. This contributed to climate change and air pollution apart from the high economic cost incurred by the industry.
Leaks occur from pressurised process equipments through bleeders, pipe connections, valves, blinds, flanges etc. And to minimise these leaks regular leak detection operations and routine inspections are required.
I worked on a project where fugitive emissions monitoring was carried out. There were 50 pairs of poles in the line of sight of gas detectors. Detectors were calibrated on Methane 2.5% LEL and 5% HEL.
These LOS detectors couldn’t detect methane gas.
Modbus 485 connector + iMX7/ AR532 IoT gateway + Det-Tronics Line of sight sensor
Detecting Hose failures
Hose failures result in asset downtime, human safety concern, repair and replacement costs. Hydraulic hose detection monitoring system detects hose failures and provides predictive maintenance notifications.
Each hose fitting has a sensor which monitors the hose condition regularly. The sensors compare the hose working in normal conditions to create a baseline which helps in comparison.
Connected Foundry in Manufacturing
Most of the equipments in small to mid sized industries require manual operation and monitoring, even for the most basic things. While remote operations of a factory is years away from mass adoption (even for a significant traction), monitoring can be, to a very large extent, can be automated.
I have visited hundreds of manufacturing units across US and there’s an interesting opportunity to leverage IoT that very few know of.
Most of the instruments in manufacturing’s foundry have an RS485 port. This RS485 protocol can easily allow anyone to capture data from it. And for any founder operations it is critical to keep track of:
- Energy consumption
- Metallics consumption
When it comes to energy consumption, here’s how operations are being carried out:
- kWh Meter manual reading is taken
- This reading is now recorded in a log book
- This information is now copied to the ERP
- Information is interpreted and feedback is delivered to the management
The table below shows the process in a bit more detail
A really long cycle to convey simpler information.
Now, let’s take a look at the state of metallics consumption here:
- Weighment done on a scale
- Again recorded in a log book
- Again entered to an ERP
- Feedback and interpretation for the work done
Here’s a diagram of a Continuous casting process unit
So, for this particular continuous casting process, we have four units:
- A weight scale
- A furnace
- A withdrawal unit
- And a cutting unit
This is a fairly simple Foundry that I visited in Nevada. Now, each of these units had an RS485 port on it. Which mean that I could simply connect another module on top of it that can retrieve information and push that information via my IoT network and bring a lot of connectivity here.
Here’s a breakup of what I did here:
- kWh meter readings from the furnace were instantly uploaded to the ERP using a hardware module that had an RS485 module
- I added an RS485 interface to weighing module and implemented batching
The end result:
- Analytics as opposed to meetings
- Alerts for calculations, per hour readings, per hour energy consumption and anomaly detection
- A wireless sensor network that reported information to a Sierra wireless gateway. Local logs were generated, edge intelligence was placed along with cellular alerts to stakeholders and decision makers
The power of IoT couldn’t be expressed in simpler terms than this, we not only were able to save this Foundry 100’s of meetings in a week, but were also able to bring workflow tracking for increased customer satisfaction as well.
Laser Engraving and Manufacturing optimization using IoT
Laser marking is extremely common in manufacturing these days. But what’s again common is the fact that most manufacturers have to spend a lot behind upkeep and maintenance of their laser engravers.
About two years ago, I visited this large facility in Nebraska that had multiple operators and PCs dedicated different laser engravers. This was inefficient and clearly violate principles of lean and six sigma.
If you don’t have this already, you should definitely consider using a solution like Lanmark LEC boards. When I introduced these boards to the manufacturing facility, we were able to bring factory automation and monitoring to laser engraves with just one PC dedicated to it.
Lanmark’s board is just an example, you could consider getting a custom board built for your organization or purchase a different one as well. The point being that these middleware type hardwares simplifies a lot of work that you could do. With Lanmark the factory was able to upload and manage laser engraving jobs (including defining 3D shapes, types, designs, etc) right from an on premise computer.
With PLCs it is extremely easy to bring in connectivity using USB and PCIexpress type interfaces and if you are savvy enough, you can get away with one on premise server and a couple of IoT gateways managing laser engraving for your factory floor.
Though, with my own experience, I didn’t liked working with a .NET based framework, takes away a lot of freedom in what you can do with IoT type open-table scenarios. But there are device manufacturers that you can pick from with your own preference.
Moulding Process Optimization and Defect Reduction using IoT
If you have ever worked with a manufacturer that does meta moulding, 90% chances are that you would already know what I am going to talk about here.
Moulding is very old, but complicated process. The biggest challenge in moulding is not about the process of how it is being carried out, but the fact that after a month of lapse when piece is worked upon, customer gets into picture and points out that you made a mistake.
Getting moulding process to become lean is often a huge challenge. Organizations worldwide have adopted a wide variety of processes to make sure they get to a lean moulding process. But, it often fails. And, when it does, there’s nothing that can be done to reduce time and material costs.
So, optimizing the quality of this process as it happens can be the key.
We can blindly go about connecting everything, but that wouldn’t even work as we have the following challenges to address:
- Internal communication lines for most manufacturers are too stressed to even accommodate anything else
- A lot of proprietary locked machines
There are some cases where I’ve seen moulding setups that can be interfaced with a gateway, so I’ll recommend an IoT solution like the one below :
If the moulding setup in your manufacturing plant has I/O capabilities, you can directly connect it to an IoT gateway in addition to your local IT systems. If that’s not possible, then you can certainly connect your local PLC setup with a Gateway that pushes this information to your cloud. A lot of your focus here would be on calibration as these are legacy systems, some of them don’t even speak on the same standards with which we were born. Finding documentations of these standards (or people who built them) is a bit difficult. At the end, you’ll end up matching bit by bit of information transfer and see if the data you fetch is right or not.
OPC UA gateway are the best fit for such environments
If you are wondering on what you would fetch with this gateway, usually most moulding setups generate CSV on a local machine. If you are fetching information directly from the PLC or sensor I/O, it would be easy to fetch raw data and transform this information to whatever you want to. In the latter, you can even create a JSON object (loved by cloud technologies) and push to it your cloud services for further processing.
Most moulding machines fetch data each millisecond during it’s process, so you would need a protocol that can transfer information as fast as possible without building an information overflow type situation. I would recommend going for OpenHART, if that’s not possible go for Zigbee or BLE. BLE would cost a lot more to you as you would end up adding to many modules to build a communication bridge.
On the cloud, I would have the following setup:
- An IoT Hub type setting to receive sensor as events or files
- Something to gather environmental data
- Custom cloud services to uncompress data (if you’ve compressed it in between) and some data pre-formatting
- Put a stream service to capture this information to be passed onto a machine learning service
- Use a Machine learning model to determine if the moulding process is going to deliver the required quality or not
That’s it, you are now lean again. With this you could be certain that there’s no way you are going to miss out on any opportunity to be lean as there would be extremely less wastage. Everyone from the customer to the management can take collective, informed and collaborated decisions.
Condition monitoring of bearing to prevent failures
We all know that any type of fault in bearing costs an unplanned downtime to the company, product quality gets at risk and safety of employees is also at stake. To avoid this either we can manually check the bearings now and then which is a very tedious task and involves life at risk or we can connect the bearings to our computers.
While working on bearing monitoring I found there vibration is such a common phenomenon which if looked upon can do wonders in case of bearings. The bearing faults which can be detected with vibration analysis are:
- Overheating – Due to overheating rings is annealed and lubricant is destroyed. Bearing capacity reduces as hardness is reduced, causing early failure
- Excessive loads
- Reverse loading – When axial load is loaded in opposite direction results in excessive stress and temperature
- Brinelling – When the load exceeds the elastic limit true brinelling occurs which is generally caused by overload
- Lubrication Failure – Insufficient lubrication results in overheating and other catastrophic failures
- Corrosion – Exposure to corrosive fluids leads to fatigue failures
- Misaligned bearings – If misaligned by more than 0.001 in./in, it can cause abnormal rise in temperature
- Irregular fits – Irregular fits are of 2 types- loose fits and tight fits. In either case there is a loss to the bearing
Piezoelectric accelerometer + Arduino + Zigbee
Condition monitoring of electrical box to prevent fires
It is very common to hear that electrical box caught fire. And to avoid this, manually checkings are conducted. It is both dangerous and time consuming.
Thermal sensors! Yes, they are used to monitor the electrical boxes, but that again requires someone to keep a note of the thermal data.
This made me curious to find a solution and help monitor electrical boxes sitting remotely.
TMP 100 + MSP430 MCU + Zigbee (It is supported by IEEE 802.15.4 and Zigbee PRO)
TMP 100 operates in the range of -55 to +125 degree C which makes it suitable for most industrial applications.
IoT with Extrusion (Impact Extrusion) processes
There’s a catch with Extrusion process machines. Most of these that I’ve personally come across have been with mid to large size manufacturing companies. And none of these companies had an off-the-shelf extrusion processing machines, they got them custom built to make sure that they can accommodate their recipes and changes that comes with each unique requirements.
The two biggest challenges that these extrusion processes face today are:
- Change of recipes and accommodating it with fluidity – Recipes can change daily, after 2 days or even weekly
- Lack of Electronic control over these recipes – They do have PLC solutions, but decision making depends upon the operation of these machines as opposed to the sea of historical information and intelligent processing algorithms
Now, these extrusion machines have PLCs from vendors like Allen Bradley and Siemens. But having a PLC shouldn’t discourage you from getting IoT into the picture.
In this case, you can simply interface these PLCs using Keyware’s IoT Gateway. This gateway also comes with a KEPServerEX where you can not only reliably carry out extrusion process automation, but can also securely store your recipes that you can manage remotely, add a ton of machine learning and AI to drive more efficiency in your extrusion process as well.
Integrating IoT with Embossing machines
As an alternative to metal die equipment, embossing machine are often a low cost alternative. Most of these Embossing machines are being used with hybrid automated/manual operations. I am not going to get into details, as most from manufacturing backgrounds would already be a lot familiar with a typical conveyor + Automated/manual shifting operations.
The way they are currently being used generates a few opportunities for IoT:
- Set Embossing parameters remotely
- Optimize per hour production capacity
Maintaining production rate while factoring in profitability and increasing your overall quality is extremely difficult.
If you look at a typical embossing machine, like the one below, you’ll get to see that it communicates information and can take responses over a RS232 port.
You already know what to, I mean, we’ve walked through thousands of other use cases where you can connect something using RS protocols.
In this case, I would make one strict recommendation as I see the need for a lot of localized decision making – Consider adding edge intelligence for defects and Quality assurance related issues here.
Since in a fully/semi-automated assembly, you would have a continuous or batch flow of metal plates for embossing operations. If something fails in one component, it is more likely to fail in rest of the batch or the other plates as well.
We usually have one operator looking at a single line of production, so for 10 lines you would need 10 people to look after it. With the IoT implementation that I’ve shown below the embossing line operations are more oriented towards reduced manual intervention and more on automated and proactive interventions.
What I personally like about RV50 is that it has an internal edge computing module that you can use in this case with minimal processing and leverage localized decision making such as alerts to managers and line operators. Upon finding that you embossing assembly may fall behind production schedule, even project stakeholders can be notified before it is too late.
RV50 fits perfectly because of Lua’s extremely proficient code that can be perfected to never fail in production environments, and the gateway is extremely rugged designed with manufacturing type environments in mind as well.
Smarter Metal bending using IoT
Like any other operation into manufacturing retaining quality with operations is extremely important. In metal bending works, it is less about carrying out automated works, but more about tracking job status and communicating this information to your MES and other systems.
Most of the large metal bending machines that you might see, like the one above, they communicate with external devices over a USB protocol/extension.
As far as the automation part is concerned, you can import job models to the machine and get newer models here.
Now, whatever facility wide protocol you have implemented, you can use that exact protocol and get a MCU based hardware unit that can fetch information from the USB port and ship it to your central wireless node.
In case if you don’t already have an existing IoT protocol that you have placed facility wide, consider implementing this hardware using Zigbee. The reason is pretty simple, there are some already existing Zigbee services that can be easily customized to create something that can be used to communicate this information.
Similar IoT implementations can be seen for:
And other operations that fall into Forming processes.
Monitoring lathe machine wear and tear using IoT
Though most modern manufacturing companies have moved on to a much more modern machining tools. But it isn’t the same for other manufacturing companies that have invested a lot 10 years ago on machining tools.
Lathe machines like these still account for more than 40% of total machining operations globally
To efficiently monitor this system, we have to place an 3 axis accelerometer on the tool holder to build a proper data acquisition.
The data thus generated often needs amplification, so we will have to Amplify and filter accelerometer’s reading for reliable and consumable information.
The overall architecture of the system could look something like this
If you have more than 20 lathe machines, using multiple channel analyzers won’t help to your connectivity needs. I would rather prefer to go with WirelessHART or Zigbee to build extended range connectivity and pool all of this data on a single IoT gateway.
While selecting a gateway to make sure that you prefer a gateway that has edge intelligence in it. If you look at what we have right here, you would find yourself getting 16 bit ADC recording from 3 channels that translates to 2KB per second. That’s 1GB of data in 18-20 hours per lathe machine.
So, with more than 20 lathe machines, you would end up pushing 30GB+ data per day to your cloud. That’s why I am recommending you to instal edge intelligence and process relevant information, and push only what’s absolutely important. Gateways from Dell, Advantech or Sierra can easily do that for you.
But this flow of information is so fast that it will exceed all in memory operations. I would recommend re-sampling this data. The frequency of sampling should be changed at 500Hz-1KHz.
To optimize the sampling rate for your machine tooling IoT, you should try to observe the data patterns. I typically plot the information we receive and then try to see how much events you can capture with a particular sampling rate.
The data should look something like this
Similar other implementations can be done for:
- IoT connected mills and milling operations
- IoT connecting Turning that includes Lathe, Facing, Boring, Spinning, Knurling, Hard turning, Cutoff
- IoT led Drilling, Reaming, countersinking, Sawing and Tapping operations
IoT in welding operations
Manage welding procedures like:
- Management of pWPS, WPQR and WPS templates as per the welding standards
- Quality control, and
- Welding management
- Welding production analysis
Earlier, welding quality depended upon the skills of welder. This in the last decade has been largely automated with the help of automated welding tools.
I don’t know why a lot of people talk about factory floors as if they are body shops that only have manual labours. That’s not true, but somehow the software world has successfully painted this image. Factories are automated, but not well connected.
Getting back to welding operations, automated machines now rule the space. But they just do what you tell them to do.
If you have sat down with a welding crew, you would know the challenges that they face. When something goes wrong in an automated welding machine, an operator has to check for the following:
- What has changed in the process?
- Has the OS been recently changed?
- What about the consumables for robotic gun?
- What about the machine parameters? Parameters like Distance, time, pulsing, peak current, BKDG time, Travel, Wirefeed, Tolerance, Weld program settings and 15 other parameters.
The problem isn’t just with the number of parameters, but also that you need to figure out the issue as soon as you can as well.
Weldeye is a remarkable organization that does this very effectively.
Using a solution like Weldeye, a welder will scan his own ID and the documents. If the welder is identified to be non-conformant, an alert will be triggered to the welding management and supervisors. Once this information is matched with the weld, a work order is generated and your welding job is locked in till it is gets a QA certification.
There is more to it. You can go on and on transforming the way traditional manufacturing shop floor works. Apart from what we have seen such as self healing machines, IoT can also enhance inventories using machine learning. Machine learning is getting integrated with IoT to make use it in supple chain and run production life cycle.
The interconnectivity along with automation is reducing human labour and time to market. This interconnectivity will give rise to another question, how important is data? Lots and lots of data will be generated which requires security through end to end transparency across the organisation.