Power, Reliability and Sustainable Design

Introduction

An Internet of Things (IoT) device is successful only when it performs its required function for the required lifetime under realistic environmental and electrical conditions. A sensor that produces accurate data but exhausts its battery in two weeks, fails during temperature variation, radiates electromagnetic interference, or cannot be repaired is not a successful engineering product.

Power, reliability and sustainability are therefore design concerns rather than final inspection activities. They influence component selection, circuit architecture, firmware, enclosure design, communications protocols, manufacturing, maintenance and end-of-life treatment. This chapter develops practical methods for estimating energy use, reducing consumption, managing batteries, using harvested energy, controlling thermal and electromagnetic risks, and designing products that remain dependable and responsible throughout their life cycle.

Learning Objectives

By the end of this chapter, the learner should be able to:

• construct an energy budget for an IoT node;
• distinguish power, energy, capacity and battery life;
• apply sleep modes and duty cycling in firmware and hardware;
• select suitable battery-management and energy-harvesting methods;
• identify thermal, electromagnetic compatibility (EMC), environmental and reliability risks;
• design for maintainability, repairability and safe electronic-waste management;
• evaluate an IoT design using technical, economic and environmental criteria.

1. Power, Energy and the IoT Design Constraint

Power is the rate at which energy is transferred. It is measured in watts (W). Energy is power accumulated over time and is measured in joules (J), watt-hours (Wh), or milliwatt-hours (mWh). The basic relationship is:

where E is energy, P is average power and t is time. This equation assumes that power is constant. If power changes with time, the general relationship is:

For a battery-powered device, average current is often more convenient than average power. If the supply voltage is approximately constant:

where I_i is the current during operating interval i, t_i is its duration, and T = Σt_i is the total cycle time. The units must be consistent, such as amperes and hours.

Average power is approximately:

where V is supply voltage. This approximation is most useful when voltage does not vary substantially and converter losses are either included in the measured current or separately accounted for.

A device may have a low average power but a high peak current. Peak current matters because batteries and voltage regulators have internal resistance and current limits. A radio transmission can briefly cause a voltage drop that resets a microcontroller even when the calculated average energy appears adequate.

Energy-budget method

An energy budget lists every operating state and its duration. Typical states include:

• deep sleep;
• sensor warm-up;
• measurement and processing;
• radio connection or transmission;
• reception and acknowledgement;
• actuator operation;
• battery-monitoring and housekeeping functions.

For each state, record current, voltage, duration per cycle and frequency of occurrence. The energy per cycle is the sum of the energy used by each state. A design margin should then be added for battery ageing, cold temperature, regulator losses, measurement uncertainty and unexpected retransmissions.

Worked example: a battery-operated environmental node

An illustrative node operates from a 3.3 V regulated supply. During a 10-minute cycle it has the following load profile:

The average current is:

I_avg = (0.020 mA × 599 s + 8 mA × 0.5 s + 120 mA × 1.5 s) / 600 s

The numerator is 11.98 mA·s + 4 mA·s + 180 mA·s = 195.98 mA·s. Therefore:

I_avg = 195.98 / 600 ≈ 0.327 mA

The daily consumption is:

Q_day = 0.327 mA × 24 h ≈ 7.85 mAh/day

A nominal 2,400 mAh battery would have an ideal runtime of:

t = 2400 / 7.85 ≈ 306 days

This is not a guarantee. If only 70% of nominal capacity is considered usable after temperature effects, ageing, conversion losses and design margin, the practical estimate is approximately 214 days. This example demonstrates why radio duration and current can dominate energy use even when the radio is active for only 0.25% of each cycle.

2. Sleep Modes, Duty Cycling and Energy-Proportional Operation

A microcontroller normally provides several power states. In active mode, the processor and peripherals operate normally. In idle mode, the processor clock may stop while selected peripherals continue. In standby or stop mode, more clocks and regulators are disabled. In deep sleep, wake-up sources are restricted and current may fall to microampere levels.

The best mode is determined by the required response time, retained memory, wake-up sources and peripheral availability. A deep-sleep mode is unsuitable if the application must respond within microseconds or continuously sample a high-bandwidth signal.

Duty cycling means operating a subsystem only for the fraction of time during which it is needed. A useful approximate duty cycle is:

where t_on is active time and T is the repetition period. If active and sleep currents are I_on and I_sleep, respectively:

This model assumes two clearly defined states and a constant supply voltage. Start-up energy, transition current and leakage should be added when they are significant.

Good duty-cycling practice includes:

1. power-gating sensors that do not need continuous operation;
2. using interrupt-driven rather than polling firmware;
3. batching measurements and transmitting one compact packet;
4. selecting communication intervals according to information value;
5. reducing unnecessary radio acknowledgements and retries;
6. disabling unused clocks, LEDs, pull-ups and interface peripherals;
7. measuring sleep current on the assembled product rather than trusting only data-sheet values.

Duty cycling can conflict with data freshness, alarm response, network synchronisation and sensor warm-up time. These requirements must be stated before selecting a sleep strategy.

3. Battery Management

Battery capacity and runtime

Battery capacity is commonly stated in ampere-hours. It is not a fixed amount of energy under all conditions. Actual capacity depends on discharge current, temperature, cut-off voltage, age, chemistry and storage history. For a battery with nominal capacity C_nom and usable fraction f, a first estimate is:

If the load has average current I_avg, the ideal operating time is:

The estimate assumes that current is expressed in the same time basis as capacity and that the battery voltage remains suitable. For a regulated system, converter efficiency η should be included. If a battery supplies average power P_b and the load requires P_load:

Battery selection must also satisfy peak-current requirements, not merely average energy requirements.

Common battery chemistries

Primary lithium cells provide high energy density and long shelf life but are not normally rechargeable. Alkaline cells are inexpensive and widely available, but their voltage and effective capacity can decline substantially at high current or low temperature. Lithium-ion and lithium-polymer batteries are rechargeable and energy-dense, but require protection against overcharge, over-discharge, excessive current and unsuitable temperature. Nickel-metal hydride cells tolerate repeated use and have robust characteristics, though their self-discharge and voltage profile differ from lithium systems.

The chemistry should be selected with regard to energy demand, peak current, service environment, charging availability, safety, transport rules and disposal arrangements.

Battery-management functions

A battery-management system may provide:

• charging-current and voltage control;
• overcharge and over-discharge protection;
• short-circuit and over-current protection;
• cell balancing in multi-cell packs;
• state-of-charge estimation;
• temperature monitoring;
• fuel-gauge communication;
• safe shutdown and event logging.

State of charge is not always equal to open-circuit voltage, especially under load or during charging. A fuel gauge may combine voltage, current integration, temperature and a learned battery model. Designers should validate estimates with the actual battery and load profile.

Power-conversion losses

A linear regulator dissipates approximately:

This applies when quiescent current is neglected. A switching converter is usually more efficient at moderate or high load, but it introduces switching noise, quiescent current and layout requirements. Efficiency is:

where η is dimensionless and normally less than 1. At very light load, quiescent current can make a converter inefficient despite an excellent data-sheet peak-efficiency value.

4. Energy Harvesting

Energy harvesting converts environmental energy into electrical energy. Solar photovoltaic, vibration, thermal-gradient and radio-frequency harvesting are common approaches.

A harvested system generally includes a transducer, rectifier or conditioning stage, energy-storage element, power-management circuit and load. Because environmental input is variable, the system must define what happens when stored energy is insufficient. Options include reducing sampling frequency, postponing transmission, degrading measurement precision or entering an emergency mode.

For a solar source, a rough energy estimate is:

where P_source is available source power, t is exposure time and η_system represents the combined conversion and storage efficiency. This is an estimate, not a guarantee, because illumination, orientation, shading and temperature vary.

A practical design compares harvested energy with consumption over a representative worst-case period, such as several cloudy days. The storage device must also tolerate the expected charge and discharge cycles. Supercapacitors offer high cycle life and high power delivery but generally lower stored energy than batteries. Rechargeable batteries provide greater energy density but age through calendar time and cycling.

5. Thermal Design and Environmental Protection

Electrical losses become heat. A simplified temperature-rise relationship is:

where ∆T is temperature rise in °C, P_loss is dissipated power in watts, and θ is thermal resistance in °C/W. This lumped model assumes steady-state conditions and a reasonably defined thermal path. For a component connected to ambient through a thermal resistance θ_JA:

where T_J is junction temperature and T_A is ambient temperature. The result must remain below the component’s permitted operating limit with a safety margin.

Heat can reduce battery life, shift sensor readings, accelerate capacitor ageing and cause processor or radio failure. Thermal design may involve larger copper areas, thermal vias, heat spreaders, ventilation, physical separation of heat sources and firmware-controlled throttling.

Environmental design considers moisture, dust, condensation, vibration, shock, corrosion, ultraviolet exposure, chemicals and temperature cycling. The enclosure, seals, connectors and cable glands must match the intended environment. A water-resistant housing is not automatically suitable for immersion, and an outdoor enclosure can overheat in direct sunlight even when ambient air is moderate.

Sensors require special attention: a protective membrane may improve ingress protection but alter response time, acoustic performance or gas diffusion. Environmental requirements should therefore be tested at system level rather than inferred only from individual component ratings.

6. Reliability Engineering for IoT Devices

Reliability is the probability that a product performs its specified function for a specified time under stated conditions. Reliability engineering begins with identifying failure modes and their effects.

A failure mode and effects analysis (FMEA) asks:

• What can fail?
• What causes the failure?
• What is its effect on the user or system?
• How can it be detected?
• How can its likelihood or impact be reduced?

Typical IoT failure modes include battery depletion, connector corrosion, memory wear, clock drift, radio interference, water ingress, firmware deadlock, sensor contamination and power-supply brownout.

Reliability is improved through derating, qualification testing, watchdog timers, brownout detection, input protection, error detection, redundant measurements where justified, safe-state behaviour and remote diagnostic logging. A watchdog is useful only if the system can recover safely; indiscriminate resets may hide a design defect.

A useful exponential approximation for constant failure rate λ is:

where R(t) is reliability at time t and λ is the constant failure rate. This model assumes statistically independent failures with a constant hazard rate. It is often inappropriate during early-life failures, wear-out, battery ageing or highly changing environmental stresses. It should therefore support—not replace—physical testing and failure analysis.

Verification should include normal operation, low-battery conditions, power interruption, repeated reset, communication loss, temperature extremes, vibration, electrostatic discharge and long-duration operation. Test records should identify hardware revision, firmware version, environmental conditions and observed failure behaviour.

7. Electromagnetic Compatibility and Signal Integrity

EMC has two complementary requirements: emissions must not disturb other equipment, and the device must tolerate electromagnetic disturbances in its operating environment. Sources include fast digital edges, switching regulators, radio transmitters, motors and poorly controlled cable currents.

Practical EMC measures include:

• short current loops in switching circuits;
• continuous and well-designed return paths;
• appropriate decoupling capacitors placed near supply pins;
• separation of noisy power and sensitive analogue paths;
• controlled cable routing and shielding where justified;
• filtering at enclosure exits;
• chassis or reference-plane design appropriate to the product;
• controlled radio and antenna layout;
• protection against electrostatic discharge at user-accessible points.

A capacitor is not an ideal device at all frequencies; its package, mounting inductance and dielectric affect performance. Similarly, a shield is effective only when termination and current-return paths are designed correctly. EMC compliance must be assessed against the regulations and standards applicable to the target market. Laboratory pre-compliance tests are valuable, but final certification remains a product-specific responsibility.

8. Maintainability, Repairability and Sustainable Design

Sustainable IoT design considers the full life cycle: extraction of materials, manufacturing, transport, use, maintenance and end-of-life treatment. Lower energy consumption during operation is important, but frequent replacement, inaccessible batteries and premature disposal can offset that benefit.

Design for maintainability means making inspection, diagnosis, cleaning, firmware updating and replacement practical. Design for repairability additionally supports disassembly and replacement of failed modules. Good practices include:

• using standard fasteners instead of permanent adhesive where possible;
• providing access to batteries and service connectors;
• using modular sensor, radio and power sections;
• documenting test points and diagnostic indicators;
• avoiding unnecessary proprietary parts;
• providing secure firmware-update mechanisms;
• recording serial numbers, revisions and service history;
• designing connectors to prevent incorrect insertion.

Trade-offs are unavoidable. A sealed enclosure may improve water resistance but hinder repair. A soldered battery may reduce size but increase disposal burden. A modular connector may simplify service but add cost, volume and possible failure points. Decisions should be justified using expected service life, safety, environment and total cost of ownership.

Electronic waste should be managed through prevention, reuse, refurbishment, material recovery and safe recycling. Batteries must not be placed in ordinary waste streams where collection rules prohibit it. Products should communicate battery chemistry, disassembly hazards and appropriate return routes. Designers must also consider data privacy: returned IoT devices may retain credentials, location history or personal data. Secure erasure and credential revocation are part of responsible end-of-life management.

9. Applied Case: Solar-Powered Agricultural Soil-Moisture Node

Background

The illustrative system monitors soil moisture and temperature in a remote field. It sends measurements through a low-power wide-area radio network and is expected to operate for several years with minimal maintenance.

Problem

An initial prototype used a continuously powered sensor and transmitted every minute. Its small battery failed during cloudy weather, and radio retries increased when the antenna became wet. The enclosure also admitted condensation.

Analysis

The design team measured actual current rather than relying only on nominal values. They found that the sensor consumed most of its energy when left powered between measurements, while frequent transmissions created unnecessary peak loads. The battery capacity was also specified at room temperature and did not represent cold-season performance.

Approach and solution

The revised design:

1. powers the soil sensor only during measurement;
2. wakes every 15 minutes instead of every minute;
3. stores measurements locally and transmits a compact summary hourly;
4. uses a regulator selected for low quiescent current;
5. adds a reservoir capacitor near the radio supply;
6. uses a rechargeable battery with protection and temperature monitoring;
7. sizes the solar panel and storage for a conservative low-light period;
8. adds a pressure-equalising vent and improves condensation control;
9. records battery voltage, reset cause and transmission retries;
10. provides a replaceable sensor module and documented field procedure.

The firmware enters deep sleep between events and postpones noncritical transmissions when stored energy is low. A watchdog and brownout detector provide recovery, while a local memory buffer prevents data loss during network outages.

Outcome

The design reduces average energy demand, avoids unnecessary radio activity and makes faults observable. It does not eliminate all risk: prolonged shading, flooding, wildlife damage and radio-network changes remain possible. However, the system now has measurable reserves, planned maintenance and a clearer end-of-life pathway.

Learning points

The case illustrates that sustainability and reliability often reinforce one another. Lower duty cycle reduces battery stress; modularity improves repairability; diagnostic data reduces unnecessary site visits; and environmental testing prevents failures that would otherwise create waste.

10. Professional Design Workflow and Best Practices

A practical workflow is:

1. define lifetime, service interval, environment and functional response time;
2. list operating states and measure current in each state;
3. calculate average and peak power, including conversion losses;
4. select energy storage with temperature, ageing and safety margins;
5. design sleep, wake-up, fault recovery and low-energy behaviour;
6. analyse thermal, EMC, mechanical and environmental risks;
7. perform FMEA and identify verification tests;
8. build prototypes that include the final enclosure and wiring;
9. test worst-case conditions and long-duration behaviour;
10. document maintenance, repair, software updates and disposal.

AI tools may assist with test-log classification, anomaly detection and energy-profile comparison, but their recommendations require engineering validation. An AI-generated battery estimate or failure prediction is not evidence unless its assumptions, data quality and uncertainty are understood.

Challenges and Limitations

Energy estimates are uncertain because battery capacity, radio conditions, temperature and user behaviour vary. Laboratory EMC results may not predict every installation. Reliability models based on constant failure rates can be misleading for ageing products. Environmental claims can also become superficial if they consider only operating energy and ignore manufacturing, transport, replacement and disposal. Consequently, measurements, conservative assumptions, traceable requirements and transparent documentation are essential.

Chapter Summary

Power-aware IoT design begins with an energy budget that includes every operating state, conversion loss, peak-current event and realistic margin. Sleep modes, duty cycling, power gating and efficient communication can greatly extend operating life, but must satisfy response-time and data-quality requirements. Battery selection requires attention to chemistry, temperature, ageing, safety, usable capacity and peak current. Energy harvesting is effective only when source variability and storage limitations are explicitly designed for.

Reliability depends on thermal control, EMC design, environmental protection, fault detection, recovery and systematic testing. Sustainable design extends beyond low power: products should be maintainable, repairable, upgradeable and responsibly managed at end of life. Battery safety, data erasure and electronic-waste regulations are important professional responsibilities. The most effective design process integrates energy, reliability and sustainability from the beginning rather than treating them as separate compliance tasks.

Key Terms

Power: Rate of energy transfer, measured in watts.

Energy: Capacity to perform work, measured in joules or watt-hours.

Energy budget: Itemised estimate of energy consumed by all operating states.

Duty cycle: Fraction of a period during which a subsystem is active.

Sleep mode: Reduced-power operating state used when full operation is unnecessary.

Peak current: Highest instantaneous or short-duration current demanded by a load.

Quiescent current: Current consumed by a device while providing little or no output activity.

Battery capacity: Charge a battery can deliver under specified conditions.

State of charge: Estimated remaining charge relative to usable battery capacity.

Power-management system: Circuitry that regulates, protects and distributes electrical power.

Energy harvesting: Conversion of environmental energy into usable electrical energy.

Thermal resistance: Temperature rise per unit of dissipated power.

EMC: Ability to emit and tolerate electromagnetic energy without unacceptable interference.

FMEA: Structured analysis of failure modes, causes, effects and controls.

Reliability: Probability of performing a required function for a specified time and condition.

Maintainability: Ease and speed with which a product can be inspected, serviced or restored.

Repairability: Ability to replace or repair failed parts rather than discard the product.

Electronic waste: Discarded electrical and electronic equipment and components.

Derating: Operating a component below its maximum rated stress to improve reliability.

Brownout: Supply-voltage reduction sufficient to cause unreliable or halted operation.

Review Questions

Short Answer Questions

1. Distinguish power from energy and give the units of each.
2. Why can peak current be important even when average current is low?
3. State two advantages and two limitations of duty cycling.
4. List four functions that may be provided by a battery-management system.
5. What are the two complementary objectives of EMC?

Long Answer Questions

1. Explain how to construct an energy budget for a battery-powered IoT node, including a suitable design margin.
2. Compare primary lithium, alkaline, lithium-ion and nickel-metal hydride batteries for IoT applications.
3. Discuss how thermal, environmental and EMC design affect IoT reliability.
4. Explain how modularity, accessible batteries and firmware support contribute to maintainability and repairability.
5. Describe a complete verification plan for a remote outdoor IoT device.

Analytical / Application Questions

1. A node sleeps at 30 µA for 899 seconds, measures at 10 mA for 0.5 seconds and transmits at 80 mA for 0.5 seconds every 15 minutes. Calculate its average current and daily consumption, ignoring conversion losses.
2. Select a power-management approach for a sensor that operates from a 4.2 V rechargeable cell but requires a stable 3.3 V supply. Compare a linear regulator with a switching converter using efficiency, quiescent current, noise and layout as criteria.
3. Design a reliability and sustainability improvement plan for an outdoor node that suffers from condensation, battery replacement difficulty and unexplained resets. Justify the tests and design changes you would introduce.

Multiple Choice Questions with Answers

Question 1

Which expression gives the average current for multiple operating states?

A. I_avg = Σ(I_i / t_i)

B. I_avg = (Σ I_i t_i) / T

C. I_avg = T / Σ(I_i t_i)

D. I_avg = Σ(I_i + t_i)

Correct Answer: B. I_avg = (Σ I_i t_i) / T

Question 2

A load draws 100 mA for 1 hour from a 3.3 V supply. Ignoring losses, what energy does it use?

A. 0.033 Wh

B. 0.33 Wh

C. 3.3 Wh

D. 33 Wh

Correct Answer: B. 0.33 Wh

Question 3

Which design decision usually reduces average energy consumption most directly?

A. Keeping the radio continuously connected

B. Increasing the LED brightness

C. Using sleep modes between measurement events

D. Adding a larger enclosure

Correct Answer: C. Using sleep modes between measurement events

Question 4

What is a major limitation of estimating battery life using nominal capacity alone?

A. Nominal capacity is independent of temperature

B. Actual capacity depends on load, temperature and ageing

C. Battery voltage never changes during discharge

D. Peak current has no effect on operation

Correct Answer: B. Actual capacity depends on load, temperature and ageing

Question 5

A linear regulator drops 5 V to 3.3 V at 0.1 A. Ignoring quiescent current, what is its approximate power loss?

A. 0.017 W

B. 0.17 W

C. 0.33 W

D. 0.50 W

Correct Answer: B. 0.17 W

Question 6

Which activity is most closely associated with FMEA?

A. Listing possible failure modes and their effects

B. Increasing radio transmit power without testing

C. Selecting a colour for the enclosure

D. Calculating only nominal battery capacity

Correct Answer: A. Listing possible failure modes and their effects

Question 7

What is the primary purpose of a watchdog timer?

A. To increase battery voltage

B. To provide automatic recovery from certain software lockups

C. To measure solar irradiance

D. To replace EMC testing

Correct Answer: B. To provide automatic recovery from certain software lockups

Question 8

Which practice best supports repairability?

A. Permanently sealing all modules with adhesive

B. Using inaccessible soldered batteries in every product

C. Providing replaceable modules and documented service access

D. Removing diagnostic information

Correct Answer: C. Providing replaceable modules and documented service access

Question 9

The relationship ∆T = P_loss θ is used to estimate what quantity?

A. Battery state of charge

B. Temperature rise caused by dissipated power

C. Radio data rate

D. Duty-cycle frequency

Correct Answer: B. Temperature rise caused by dissipated power

Question 10

Which statement correctly describes EMC?

A. It concerns only the device’s energy efficiency

B. It concerns both electromagnetic emissions and immunity

C. It applies only to battery chemistry

D. It eliminates the need for environmental testing

Correct Answer: B. It concerns both electromagnetic emissions and immunity