IoT and Field Integration | Bringing Production, Warehouse, and Field Data to Management

Do you learn about machine failures on the production line only after they occur? Do you experience product loss when warehouse temperatures exceed critical levels? Does your field team work with paper forms, with data reaching the central office two days later?

In this guide, you will learn how to transfer production, warehouse, and field data to your central management systems in real-time using IoT (Internet of Things) technology, which sensors to use, how to design the connectivity architecture, and how to operate during internet outages.

What is IoT and Field Integration?

IoT transforms the physical world into digital data

IoT (Internet of Things) is the technology that enables physical devices (machines, sensors, vehicles) to send and receive data over the internet. Field integration is the process of transferring this data to central systems such as ERP, BI, or MES and converting it into business decisions.

In the traditional approach, a field employee reads a machine counter twice a day, writes it down, and enters it into Excel in the evening. With IoT, a sensor takes automatic readings every 5 seconds, data flows to the system instantly, and management is notified if there’s an anomaly.

Key Components of IoT Integration

• Sensors: Measure physical quantities (temperature, pressure, vibration)
• Gateway: Collects sensor data, performs protocol conversion, and stores data locally even without internet
• Cloud/Server: Stores data, performs analysis, generates reports
• API/Integration: Enables data exchange with systems like ERP, MES, dashboards

Example: A vibration sensor on the production line detects that a machine has gone outside its normal operating range. The gateway adds a timestamp to the data, sends it to the cloud, and the system automatically creates a work order for the maintenance team. The entire process is completed within 30 seconds.

Operational Benefits: Real-Time Monitoring

IoT dashboards provide instant visibility

1. Elimination of Manual Data Entry

Data that field employees used to write down and bring to the office is now automatically transferred to the system by sensors. Human error is reduced to zero, and data latency drops to minutes.

2. Predictive Maintenance Alerts

Detects anomalies in vibration, temperature, and current values before a machine fails. Scheduled maintenance is performed, minimizing downtime.

3. Inventory and Logistics Visibility

Real-time monitoring of warehouse temperature, product location (RFID), and vehicle location (GPS). Cold chain violations are detected instantly, and vehicle routes are optimized.

4. Increased Decision-Making Speed

Senior management views current data on live dashboards. Instead of asking “What happened last week?”, they get answers to “What is happening right now?”. Strategic decisions become data-driven.

5. Energy and Cost Optimization

Machine-specific energy consumption is monitored, and idle equipment is identified. Energy costs are proportionally reduced.

Sensor Types and Use Cases

Each sensor type measures a different physical quantity

1. Temperature and Humidity Sensors

Use Case: Warehouse climate control, production environment, food cold chain.
Example: Automatic alarm when warehouse temperature exceeds 8°C.

2. Vibration Sensors (Accelerometer, Vibration)

Use Case: Machine health monitoring, maintenance planning.
Example: Notification to the maintenance team when vibration levels on a CNC machine exceed normal limits.

3. Current and Energy Sensors

Use Case: Energy consumption monitoring, idle equipment detection.
Example: Compressor running idle overnight, detecting energy waste.

4. Pressure Sensors

Use Case: Hydraulic and pneumatic systems, liquid tanks.
Example: Oil level warning when pressure drops in a hydraulic press.

5. GPS and Location Sensors

Use Case: Vehicle tracking, field team routes, logistics optimization.
Example: Which customer the service team is at, how long ago they left, traffic conditions.

6. RFID (Radio Frequency Identification)

Use Case: Product tracking, inventory counting, entry-exit control.
Example: Every pallet passing through the warehouse door is automatically recorded in the system.

7. Camera and Image Sensors

Use Case: Quality control, AI-assisted defect detection.
Example: Surface defects on the production line are automatically detected, and defective products are separated.

Connectivity Architecture: Edge, Gateway, and Cloud

Edge, Gateway, and Cloud form a three-layer architecture

Layer 1: Edge Computing

The layer closest to the sensor. When very fast response is needed (e.g., emergency stop), data is not sent to the cloud; the edge device makes the decision directly. Latency is at the millisecond level.

Example: When machine vibration reaches a dangerous level, the edge device cuts the relay without consulting the cloud, stopping the machine.

Layer 2: Gateway (Data Collector)

Collects data from multiple sensors, performs protocol conversion (like Modbus, OPC UA, MQTT), packages the data, and sends it to the cloud. Its most important feature is the offline buffer. When the internet connection is lost, it stores data in local memory and automatically uploads it when the connection is restored.

Example: There are 12 machines on the production floor, each using a different protocol. The gateway collects them all and sends them to the central system via HTTPS.

Layer 3: Cloud (Cloud Server)

The layer where all data is collected, analytics are performed, and reports are generated. ERP, BI, and dashboard integrations occur at this layer.

Example: Production data from all branches is consolidated in the cloud, and management monitors the entire operation from a single dashboard.

Architecture Selection Criteria

• Decision Speed: If milliseconds are required, use edge; if minutes are sufficient, use cloud.
• Internet Reliability: If low, a gateway is mandatory.
• Data Volume: If high, perform pre-processing at the edge and then send summaries to the cloud.
• Security: If critical data is involved, edge/gateway encryption is essential.

Data Collection and Processing Layers

IoT data goes through several layers from raw sensor readings to business intelligence reports:

Layer 1: Raw Data

The direct output from the sensor. Example: “43.2°C, 12:34:56”

Layer 2: Cleaning and Filtering

Erroneous readings (outliers) are filtered, and unit conversions are performed.

Layer 3: Contextualization

Metadata is added to the data: which machine, which location, which process.

Layer 4: Aggregation and Analysis

Hourly and daily averages are calculated. Trends are identified.

Layer 5: Business Rules and Actions

When a threshold is exceeded, a work order is automatically created, or a notification is sent.

Layer 6: Reporting and Visualization

Presented as dashboards, Power BI reports, or Excel reports.

Example Data Flow:
Sensor: 92.4 mA → Gateway: “Machine-003, 92.4 mA, 14:22:10” → Cloud: “Normal range 80–100 mA, OK” → Dashboard: “Machine-003 is healthy, consumption is normal”

Turkey/TRNC Realities: Offline Mode and Internet Outages

Internet outages are a reality in industrial zones in Turkey and across the island in TRNC. The success of an IoT project depends on its ability to operate without internet.

Offline Mode Architecture

The gateway device stores sensor data in its local memory (SD card, SSD). Automatic synchronization begins when the internet connection is restored. The user does not need to perform any manual intervention.

Offline Buffer Capacity Calculation

Example calculation:
– 10 sensors, each reading every 10 seconds (60 readings per minute)
– Each record is 100 bytes
– Hourly data: 60 × 60 × 100 bytes = 360 KB
– 7-day buffer: 360 KB × 24 × 7 = 60 MB

Result: 1 GB of memory can handle over 4 months of outages. In reality, the longest outages are usually a few hours, ensuring no data loss.

TRNC Specific Situation

Due to frequent power outages in TRNC, a UPS (uninterruptible power supply) is mandatory for gateway devices. A battery capacity of at least 2 hours is recommended.

Offline Mode Critical Points

• Time Synchronization: The gateway’s clock must be accurate (NTP server).
• Memory Management: If the buffer fills up, the oldest data is deleted (FIFO).
• Conflict Management: The same data should not be sent twice (unique ID).
• Error Notification: If there has been no internet for 24 hours, an SMS is sent to management.

Field Example: Multi-Location Production and Warehouse

7 Common IoT Implementation Mistakes

1. Starting Without Offline Mode

Ignoring the reality of internet outages. Installing without considering gateway buffer capacity. Result: data loss during the first outage, eroding user trust.

2. Too Many Sensors, Too Little Analysis

The “let’s put sensors everywhere” approach. Data is collected, but no one analyzes it or makes decisions. IoT should generate actions, not just data.

3. Neglecting Security and Encryption

Sensor data sent over open protocols (HTTP, plain MQTT). Production data can leak externally, or devices can be hijacked. TLS/SSL is essential.

4. Forgetting Time Synchronization

If the gateway clock is incorrect, data timestamps will be wrong, leading to inconsistencies in reports. Automatic time synchronization via an NTP server is mandatory.

5. Not Planning for Scalability

Starting with 3 pilot sensors, but the gateway capacity is insufficient when scaling to 100 sensors. A scalable architecture should be chosen from the outset.

6. Skipping Maintenance and Calibration

Sensors drift over time (measurement shifts). If calibration is not performed every 6 months, data reliability decreases. A maintenance schedule is essential.

7. Delaying ERP Integration

The common approach is “Let’s install the sensor first, integration later.” Data exists on the dashboard but is not integrated into business processes. The benefit of IoT data is limited if it’s not transferred to the ERP.

Prevent mistakes with proper planning

IoT Project KPI Table

Metrics to track to measure the success of IoT integration:

IoT Integration Checklist

Frequently Asked Questions (FAQ)