Machine Dossier – Industrial IoT Plant Reliability Ecosystem
The mother company of our client – a Japanese company specializing in solutions which helps to assure smooth, stoppage-free plant operation – has entrusted its US-based subsidiary with the development of a new Industrial IoT (Industrial Internet of Things) solution which would measure vibrations of various machines, process the collected data, provide feedback to technicians and occasionally send alerts in case of emergency (you would be surprised how much information is contained in vibrations).
After struggling to find a competent IoT software development partner in the USA for almost 6 months, our client has finally turned to us. As you may have already found out, a lot of folks claim they know how to approach complex IoT projects, but do not necessarily have hands-on experience. They start to code without prior preparation and analysis, and thus fail to cover many important aspects. Security and future scalability of the IIoT solution are some of the prime victims of such approach.
Until recently our client (and its clients) had to send engineers who would manually inspect the condition of the machines. What is worth mentioning, these are not some basic appliances but industrial grade/scale manufacturing machines which should be serviced by adequate professionals. The whole process of maintaining such equipment tends to be cumbersome and thus pricey. It also involves a great deal of scrupulous planning, scheduling and possible interference in the manufacturing process due to partial shutdowns at the time of inspections. Oftentimes locations of those facilities are remote and not easily accessible so if something unexpected happens it can take more than a few days for the engineers to arrive and provide a fix. If you are in manufacturing business, you know that any potential stoppage could quickly lead not only to delays and frustration but also cost you a fortune.
What we deliver to our partners is not just software. Anyone could do that. As a true R&D partner, our goal is to craft innovative solutions which deliver tangible values to our clients. With Machine Dossier, we started by defining the exact concept of the solution, what should be achieved, how the value for the users will be generated, what is a must and what it would be nice to have. We talked about safety requirements, the type of supporting infrastructure, the pros and cons of particular solutions and a ton of other aspects. We call this phase PROVIDING A SOLUTION and in this particular case it took about 3 weeks. This step can have a tremendous effect in the future as it involves a lot of strategic decisions, so it just has to be done diligently.
For Machine Dossier, we, together with our client, decided to use Amazon Web Services (AWS) as a cloud platform provider. It is versatile, gives you plenty of options in the form of ready to use services, virtual machines etc. It is also highly scalable: dirty cheap to start with and very easy to scale up or down “on the fly”, which helps reducing total cost of the infrastructure substantially. If it is not a must for some specific reasons, why invest in your own servers, which have a good chance to sit idle for most of the time and still cost you money, if you can pay-as-you-go? We are 100% vendor agnostic – if we recommend you anything, we only have your best interest in mind and are always able to explain the benefits and potential drawbacks in detail.
After making all those initial decisions we entered RESEARCH AND DEVELOPMENT phase. We have been responsible for back-end (part of the system which is invisible to users where all the logic resides), front-end (part of the application the user interacts with directly) and UI/UX (User Interface and User Experience) developments as well as setting up all the cloud infrastructure. Our client, as a specialist in hardware, was in charge of the development of a wireless vibration sensor, which would be a data collection component of this ecosystem.
We used Agile/ Maturity approach to Project Management. 2 week sprints allowed flexibility and an iterative approach to development and were a part of a longer, 3 months’ cycle during which we delivered agreed-upon, fully functional and tested Product releases. Due to the complexity of IoT solutions, in most cases it is not that effective to stick to just one methodology. To get the best results, you need to take what is best from both worlds.
RESEARCH AND DEVELOPMENT took altogether 9 months of intensive work. Team roles included all the required positios, such as: Project Manager, Business Analyst, UI/UX Designer, Architect, Java Developers, Quality Assurance and finally Configuration Manager.
The idea behind Machine Dossier is very simple. It’s about saving time and money.
After less than a year of RESEARCH AND DEVELOPMENT we went live and began SERVICE DELIVERY phase. During that period Pro4People was responsible for maintaining the system as well as the second and third line of technical support. It was also an important time for us, as we were able to collect feedback from the users directly. It helped us to finetune and adjust future developments and releases of Machine Dossier platform.
You can have the best system, but if it cannot be trusted or, even worse, accessed, it will be quickly stopped being used. The solution we created was accessible 99,76% of time during the first 2 years of operation, confirming the quality of delivered work and the reliability of the solution as a whole.
Nowadays, using Plant Reliability Platform – Machine Dossier, all the machines are automatically inspected from the headquarters of the company in Japan. The sensor which was designed and used here is the so-called wireless intelligent edge device, meaning that it is basically a tiny computer capable of doing initial computations at the place, prior to sending them to the cloud. It limits the number of data uploaded to the servers, which can have a huge impact if you consider the complexity of data, numerous worldwide locations, including remote areas where access to the Internet can be limited or pricey, as well as millions of potentially connected sensors.
Users can setup the precise schedules for collecting data and thus do it on a daily basis, rather than every few weeks or months. The solution also does not require human attention unless it detects some abnormalities. It alerts then the relevant people responsible for particular machines, collects additional data, so that they can get full insight into the situation and react accordingly. Additionally, they can share the data with fellow specialists so that they can exchange knowledge and work collectively.
Before our client launched this product, engineers were sent for inspections to report most of the time that everything was “OK”. Just imagine the number of saved hours, business trips and money!