Using Artificial Intelligence and Internet of Things to Improve the Efficiency of HVAC

IoT will help reduce operation costs of HVAC

HVAC is the most energy-consuming system in homes, accounting for nearly 50% of the total energy consumption of buildings. Therefore, adopting an efficient HVAC system will significantly reduce the use of building energy. The use of manually controlled chillers or heater system usually leads to energy waste. In order to solve these problems, the smart HVAC system with artificial intelligence that can be used in conjunction with various IoT sensors will optimize the operation of HVAC and achieve remarkable energy saving efficiency.

"IoT" is a technology that connects various devices to the Internet. This system relies on data collection, which is used for monitoring, controlling and transmitting information to other devices through the Internet and different wireless technologies (including Zigbee and BLE). Combined with the judgment of artificial intelligence, IoT devices can automatically activate specific operations when specific situations occur. Because these IoT devices are interconnected, it is easier to manage multiple operations. In fact, IoT home devices will help reduce costs and energy, and save valuable time for users.

Facing the huge demand in the market, the AI IoT business development team of Arrow ESC has developed an overall solution that provides the market with key functions such as low power consumption, small size, high security, intelligence (artificial intelligence capability) and cost-effectiveness.

Highly integrated and cost-effective complete system

STMicro’s STM32WB55 wireless MCU is used in the control part of IoT and HVAC system, which is compatible with various protocols from point-to-point to mesh network, support Bluetooth^® LE 5.2, Bluetooth Mesh, OpenThread, Zigbee 3.0, and the open 2.4GHz wireless standard, and has excellent low current consumption with built-in security function. The peripherals supported by STM32WB55 include I²C, USART, LP-UART, SPI, USB 2.0 FS device supporting battery charge detection, SAI, Quad-SPI, 16-bit timers (including LPWM and low-power timers). It can operate in a voltage range of 1.7 to 3.6 V (DC/DC, LDO), and a temperature range of -40°C to +105°C.

The STM32WB wireless MCU has dual-core M0+M4 architecture processors and radio transceiver, combined in a deeply integrated and cost-effective system on chip. Therefore, the STM32WB solution is a perfect match between real-time and/or energy-efficient applications.

In respect of temperature sensing, STMicro’s digital ambient temperature sensor STTS75 is adopted, which has a low power consumption of 75 μA and a resolution of up to 12 bits, and can digitize the temperature with a resolution of up to 0.0625℃, so as to ensure high accuracy in the whole working range from -55℃ to+125℃. Other functions of this solution include one-shot temperature reading with I2C/SMBus output and energy saving functions with selectable addresses (up to 8), dedicated pins with interrupt or thermostat/comparator functions, and programmable threshold with hysteresis function, etc.

Hardware design with low power consumption and long battery operation

The hardware development of this solution is designed for low power consumption and battery operation, which target is for 24 months operating with only 2 AA batteries. In order to achieve this goal, a variety of low-power components must be used, including STMicro's digital temperature sensor with 75 μA low power and up to 12-bit resolution, and STMicro's dual-core RF MCU with 2.1 μA power-saving mode current and 5 μs fast wake-up capability. In addition, a Zigbee SED (Sleepy End Device) that performs RTC wake-up in Stop 2 mode (2.1 μA) and runs in a 6-second polling method between host and client synchronization is required.

Moreover, Murata IRA-S200ST01A01 PIR sensor with Rohm BD9251FV amplifier solution is also used. This is a cost-effective PIR solution, which consists of Rohm BD9251FV amplifier and Murata 2 cells PIR sensor. Rohm BD9251FV integrates an amplifier for sensor output, a comparator for sensor output, and built-in voltage regulator and moving detector.  It can provide digital output with direction information (left->right)/(right->left).

In addition, it is matched with ST VL53L5 ToF 8x8 multizone ranging sensor, which has multizone ranging output of 4x4 or 8x8 separate zones, and can support ranging up to 400 cm. By default, the VL53L5 is in low power mode, consuming only 45 μA power. By using PIR in the system, it will wake up the TOF sensor when a human body is detected. The 8x8 TOF sensor can be used to reduce false triggering and estimate the position and moving direction of the human. If the detected human is only moving in a specified direction, artificial intelligence (AI) can be added by sending only notifications, thus further reducing power consumption to reduce sending unwanted notifications.

Support OTA firmware upgrade across protocol stacks and applications

In respect of software development, it can support over-the-air (OTA) firmware upgrade across STM32WB M0 (Zigbee stack) and M4 (application) and the host MCU STM32L4 (UI FW).

This solution has independent radio activity capability, can upload data to mesh network or smart phone, supports OTA of radio protocol stack or application firmware, can run on Arm Cortex-M0+, has energy saving mode, consumes only 2.1 µA when RAM and RTC run, and can wake up quickly at 5 μs speed. The main application activities include computing data (sensor fusion, etc.), having a flexible Arm Cortex-M4 CPU speed of up to 64 MHz, supporting the batch acquisition mode (BAM) with CPU and flash memory turned off, having dual CPU activity, and the power consumption is only 50 μA/MHz. Radio and applications can run independently, supporting super power saving mode, and the power consumption is less than 50 nA when shutting down, which can save battery energy.

In terms of power saving function, the STM32WB device has FlexPowerControl function, which can improve the flexibility of power mode management and further reduce the total application power consumption. The operation mode can support system clocks up to 64 MHz, and the power consumption is only 117 μA/MHz. The STM32WB device supports eight main low power modes, including low power run, sleep, low power sleep, stop 0, stop 1, stop 2, standby with RAM retention, and SHUTDOWN mode with standby and shutdown. Each mode can be configured in many ways, providing several additional sub-modes.

The solution also uses the STM32WB built-in switch mode power supplier (SMPS) to power the digital core and radio LDOs. SMPS supports switching on the fly. When the VDD power supply is drops below 2.1 V, it will automatically switch to bypass mode. SMPS operation mode (on and off) will follow device mode. In order to remove any noise from SMPS during ADC conversion, software may switch SMPS mode on the fly.

In the future, this solution will further develop STM32WB wireless MCU to support 160x128 dots TFT LCDM with SPI interface, or use -Host MCU STM32L4x with FMC (Flexible Memory Controller) to support TFT LCDM with 8080 16-bit parallel interface. it will also be combined with ST’s MEMS microphone MP34DT06, which can transmit audio stream through BLE with STM32WB. In addition, it can be combined with Murata Type 1YN Wi-Fi module (CYW43439) as Wi-Fi component of BLE/Zigbee gateway solution.

Conclusion

Using AI and IoT technology can effectively improve the operational efficiency of HVAC system. The BLE/Zigbee IoT solution designed by Arrow Asia Pacific in cooperation with STMicroelectronics will provide all customers with low-power and cost-effective solutions. Customers can use this hardware and software reference design to speed up product development and greatly shorten the time to market. Interested customers can contact Arrow for more details.