How IoT Revolutionizes the Manufacturing Industry: Enhancing Production with Data-Driven Insights

Peter Nachtwey posed an intriguing question: What steps would you take? Which data points would you prioritize gathering, and in what ways would you utilize that data? Having participated in various data collection initiatives, I've noticed a recurring challenge: the volume of data often surpasses its utility. Organizations frequently amass vast quantities of data without a clear strategy for its application, leading to overcrowded storage. The crux of the issue lies in the proper definition of data, which directly influences its ability to deliver a return on investment. While gathering data is relatively straightforward, extracting actionable insights is considerably more complex. Here are some practical applications for data insights: 1. Optimizing resource distribution between facilities 2. Benchmarking production outputs across different plants 3. Analyzing shift performance variations 4. Identifying production bottlenecks It's crucial to ask the right questions: Why are resources misallocated? What causes production disparities? Where do bottlenecks occur? Expanding on this, Peter Nachtwey’s reflections hit the mark. My experiences in data collection projects have underscored the importance of not just collecting data but effectively leveraging it. A data scientist's role, for example, is to convert raw data into meaningful predictions and solutions. Consider a retail scenario where data shows 20 pairs of shoes and 15 pairs of socks sold on a particular day. These numbers, although simple, can reveal purchasing trends, such as how many shoppers bought shoes and socks together, and whether it would be advantageous to display socks alongside shoes, based on customer habits. However, when viewed as mere figures on a spreadsheet, they appear as basic sales statistics—20 shoes and 15 socks, with correlated profits. Thus, the process of collecting and analyzing data is undeniably challenging. Furthermore, understanding operator behavior on the production line is equally significant. For instance, monitoring how frequently an operator initiates machine cycles or noting the time when work commences after a shift starts is vital—especially if a shift begins at 7 AM but the first task isn't initiated until 7:20 AM. In summary, the collection and forecasting of data should align with specific objectives and the unique demands of the production cycle. I appreciate your time in reading and sharing your thoughts! Thank you!

• 24-03-2025
• backendcode

Your main challenge will likely be collaborating with the local IT department. While smaller companies might be more manageable, they often lack substantial financial resources. I've added extra code to machines to calculate and track production metrics, but guess how often these insights are utilized? Never are they discussed in production meetings, like when someone might say, "The data from the HMI concerned me about the last shift." Instead, issues are logged on paper and then filed away in cabinets, rarely to be revisited. Perhaps you could offer a valuable service by enhancing the accessibility and usability of their existing data. In my view, maintenance reminders on machines are somewhat outdated and irritating. However, being alerted when a sensor reading has been fluctuating recently is beneficial. Similarly, knowing that a motor is drawing more amps than usual is helpful. Yes, I understand the importance of greasing bearings and changing air filters, but it's hard to perform preventive maintenance on an operational machine. Most of the equipment I deal with operates on a run-to-failure strategy.

• 24-03-2025
• seth350

Seize the moment! Today is your opportunity to capitalize on the latest trending buzzwords and turn them into profit. Don't wait for tomorrow. Although there might not be groundbreaking innovations or significant improvements, this is how it has traditionally been in the industry.

• 24-03-2025
• spaderkung

backendcode commented: "Hey cardosocea, Capturing crucial information from the production line is paramount. Leveraging this data effectively can significantly benefit us in the future. For instance: 1. **Predictive Maintenance**: By calculating optimal times for routine upkeep, we can prevent machine breakdowns. While some machines aren't serviced if they're operating well, maintaining records of regular maintenance can extend a machine's lifespan. This data becomes invaluable over time and can trigger reminders for scheduled servicing. 2. **Inventory Management**: At my company, Canada’s leading automotive firm, production delays often stem from raw material shortages. IoT can transform this process by reliably tracking raw material availability. Although there may be paper records, knowing the exact quantity used on specific production lines enhances inventory accuracy. This systemic record-keeping provides substantial insights. 3. **Quality Control and Reporting**: Meeting production goals is great, but logging data on parts that fail quality tests or require rework is equally essential. Detailed records drive improvements. Data collection specifications might vary across production lines, but standardization can complement tailored approaches. Applying Big Data tools enriches our capacity to predict and refine outcomes. Major corporations employ data scientists to decipher customer behaviors, underscoring that harnessing and applying accurate data is crucial in IIoT. However, selecting the appropriate algorithms and predictive tools poses a challenge, albeit a transformative one for this industry. Your thoughts? Thanks for reading." Notice: "Hello, perhaps my message was unclear. I acknowledge your valid points. However, my concern lies in the way companies push IoT innovations. Competent personnel with a robust data historian can usually achieve the same outcomes. Why should a company invest in a generic, expensive, internet-dependent tool when it already possesses specialized data analysis resources? Data collection usually suffers from poorly placed sensors or inadequate inquiry. Departments within a plant often have distinct data needs—production may disregard alarms that are crucial to process safety and maintenance. Similarly, maintenance is more focused on motor currents than other departments. Ultimately, if IoT adoption is needed to align decision-making with data, it's a step forward. However, true benefits will arise only after companies shift their mindset, not just their tools."

• 24-03-2025
• cardosocea

To effectively drive engagement and understanding, it's essential to demonstrate the possibilities rather than simply articulate them. I spearheaded a pioneering initiative to integrate IoT technology (even before it had that name!) into one of our major products. Initially, I repeatedly informed the Tech Support team about the extensive capabilities available, but their interest was minimal. They couldn't grasp the potential of that vast data. Their focus was solely on receiving familiar data via modem connections. However, once my team showcased the data in practical applications, they quickly recognized its immense value and couldn't imagine managing tasks without it. A similar scenario unfolded with a manufacturing engineer I knew, who required specific production line metrics. He developed a straightforward VB app to collect the data and displayed it prominently on a large screen near his desk. With visibility into what could be achieved, management was enthused and eager for more insights. Simply claiming data collection capabilities is ineffective until stakeholders witness its tangible benefits firsthand. The challenge isn't technological; it's about persuasion and demonstration. Build a pilot project, and let the results speak for themselves. Once people see what's possible, they truly understand and embrace the potential.

• 24-03-2025
• Lyndon A

Lyndon A remarked: "It's essential to demonstrate options to people rather than just telling them. I spearheaded a project integrating IoT technology into one of our company's major products, even before the term 'IoT' was coined. Initially, I tried communicating the possibilities to our Tech Support team, but they weren't interested. They couldn't fathom what they would do with all that data, content with the limited insights gained through modem connections. It wasn't until my team showcased the data visually that they recognized its value. Once they experienced its benefits firsthand, they couldn't imagine functioning without it! Similarly, I knew a manufacturing engineer who needed specific production line metrics, so he developed a simple VB application to display this data prominently in his workspace. When management saw the potential, they embraced it eagerly. Simply stating data can be gathered is ineffective until its practical utility is demonstrated. The issue isn't technological but rather in the presentation—it's a sales challenge. Instead of verbal explanations, create a tangible pilot project. Once people witness the possibilities, they get it. I echo these sentiments completely. Most of my implemented ideas weren't proposed in initial meetings or brainstorming sessions. Instead, I created solutions and presented them, leading to widespread adoption and requests for similar innovations in future projects. While it’s easy to provide exactly what’s requested, a great engineer proactively brings fresh ideas to life, even while fulfilling existing responsibilities. This post served as a valuable reminder for me as well. Excellent insight, OP."

• 24-03-2025
• alive15

backendcode shared some insights: I'm planning to launch a company that focuses on implementing IoT technology to streamline production lines and assist integrators in setting up systems designed to gather valuable production data and information for decision-makers. However, your approach to data collection might alarm upper management. The reason is that once databases are linked to the Internet or stored on the cloud, they could potentially become accessible to parties who may not have your company's best interests at heart. Industrial floors first adopted data collection over thirty years ago, and since then, most serious businesses have operated some form of Manufacturing Execution System (MES) database to manage Production, Efficiency, and Quality. PI (Process Information) databases record time-series data, each with a timestamp, whereas Relational databases, such as SQL databases, store data in user-defined, easily-retrievable structures (rows and columns). Major industrial software providers typically offer solutions for both of these data collection methods. The key difference between traditional data collection methods and your proposed implementation lies in the location and operation of the databases. Existing databases operate within highly secure, proprietary systems, distinctly separated from public networks. In my opinion, highly sensitive IoT data collection without proprietary safeguards is not advisable anytime soon, as data transmitted over the Internet or web remains vulnerable to unauthorized access. When presenting your proposal, focus on the benefits of data collection, and assure stakeholders that databases will remain within secure, high-availability company infrastructure. It’s crucial to avoid mentioning terms like 'Internet,' 'Cloud,' or 'Web' to prevent raising concerns about data security.

• 24-03-2025
• dmargineau

Our comprehensive system ensures seamless communication through a fault-tolerant fiber-optic network. We rely on CAT 5 cables extending from our switches, ensuring reliable connectivity. If the fiber link between Switch 2 and Switch 3 experiences an issue, an alternative fiber connection from another switch ensures continuous communication. Our system leverages SQL for interfacing PLCs with our reporting system, effectively channeling data into designated tables. A report generator compiles this information into a structured format at predetermined intervals. As PLC programmers based in the IT department, we have remote access to nearly every PLC in the facility, enabling us to promptly address operator challenges visible on all connected PCs. Important Note: Maintain two distinct networks at all times. Network #1 caters to corporate and management activities, providing access to data generated from reports. Network #2 is specifically for PLC operations, with restricted management access. This separation, secured by a firewall, prevents unauthorized interference with PLC programs, such as altering specifications and timers without understanding the outcomes or current machine operations. Although we have remote access capabilities, any program modifications follow strict protocols to ensure transparency when implementing changes.

• 24-03-2025
• James Mcquade

dmargineau mentioned that your suggested approach for data collection might seriously unsettle upper management. The main concern is that once these databases are connected to the Internet or stored in the cloud, they could become accessible to external parties, posing potential risks to the company's privacy and security. Moreover, some key executives, who should play a significant role in the project, might resist this change due to apprehensions about data scrutiny and accountability. James Mcquade advised exercising caution by maintaining two distinct networks: first, a corporate network for management to access and analyze data reports, and second, a PLC network isolated from management access. However, I would suggest implementing three separate networks for optimal security and efficiency: corporate, SCADA, and PLC networks. The specific configuration would depend on your infrastructure's layout and requirements.

• 24-03-2025
• cardosocea

cardosocea mentioned: Furthermore, certain key project stakeholders might not appreciate their managers sifting through the data and challenging their decisions. Expand for more... Workplace politics, anyone?!! This is a highly relevant observation nonetheless.

• 24-03-2025
• dmargineau