I. Position of This Article

This article explains under what constraints and design choices off-grid was realized as an actual system. In Part 1, we defined off-grid as a design principle. In Part 2, we evaluated energy systems as control structures and showed that off-grid naturally emerges as the optimal solution.

Here, we整理 how that conclusion was implemented: not in terms of individual components, but as a system-level design that had to satisfy concrete constraints while remaining operable over long time scales.

II. Historical Background and Initial Conditions

When we began validating the off-grid systems that would later become the prototype for Personal Energy® around 2008, the current technological and social premises did not yet exist.

Lithium-ion batteries were known in laboratories and theoretical studies to offer high performance, but in terms of real-world use, they had not yet achieved:

  • large capacity,
  • mass production at scale,
  • or proven reliability for industrial applications.

Similarly, technology capable of measuring and controlling large currents or high voltages at millisecond resolution was, outside of research environments, not yet realistically deployable.

The same was true for solar PV. This was just before “mega-solar” plants began to appear in the market, and just before decarbonization became a widespread social “movement.”

The market at that time favored partial optimization over global optimization, and short-term diffusion and volume over long-term operation. The perspective of designing energy as a system was still weak. Under such conditions, attempting to implement an off-grid system premised on long-term continuous operation was in no way an obvious or easy choice.

III. Constraint Conditions in Off-Grid Implementation

In designing this system, what we first defined was not an “ideal,” but a set of constraints. The explicit constraint conditions were as follows:

  • Do not assume synchronism with the utility grid.
  • Do not assume external control.
  • Do not assume permanent on-site operators.
  • Do not separate “emergency” and “normal” modes.
  • Must withstand long-term continuous operation.

These were not ideological requirements. They were necessary conditions for making real-world operation possible.

IV. Definition as an Autonomous Distributed System

There are multiple definitions of autonomous distributed systems. A concise and appropriate one is:

“A structure in which each component of a system behaves autonomously, and through mutual coordination among components, the system as a whole generates order.”

This concept was proposed in the late 1970s and, from the 1980s onward, has been applied to real-world systems in fields such as railways, steelmaking, and industrial control. In other words, autonomous distributed systems are not a brand-new idea; they are a mature concept that has been implemented from the beginning as system-of-systems.

When applying this to electric power and energy infrastructure, three implementation requirements become indispensable:

  • What is used as the primary power source?
  • Can autonomous operation and fail-safe behavior be maintained independently of internal and external disturbances?
  • Can the system withstand 24/7 operation over time spans of 10 years or more?

Our system was designed on the premise that all of these conditions must be satisfied.

V. Overall System Architecture

At an abstract level, the system architecture can be organized into the following layers:

  • Primary Power Source
  • Energy Buffer (storage elements)
  • Load-Side Control (load priority)
  • Local Autonomous Control (control system)

The crucial point is that there is no central controller. To implement off-grid, autonomy, distribution, and asynchrony in the energy domain, the resulting structure is analogous to the historical shift from office computers (host systems) to personal computers in information processing.

From this design philosophy, the development codename naturally became “Personal Energy.”

VI. Primary Power and the Logic of Load Control

For the primary power source, we chose solar PV, which is available almost anywhere on Earth and relatively less exposed to geopolitical risk.

The instability introduced by weather and seasons was assumed from the design stage. We leveraged nationwide empirical data obtained via our own smart meters since 2010 to model the actual generation profile.

To smooth this structural imbalance, we placed batteries as an energy buffer and adopted a hierarchical architecture based on chemical characteristics, consisting of High / Middle / Low tiers.

On the load side, instead of simply following demand, we implemented what we call “sufficiency-aware” control. The key is not prediction. It is to observe the state and then steer the system so that it converges within the given constraints.

All load-side protocols are absorbed within the off-grid layer. Dependence on any specific equipment vendor or communication protocol is deliberately eliminated.

VII. Control Philosophy: Autonomous, Distributed, Asynchronous

The control philosophy of this system can be summarized in three principles:

Autonomous

State transitions are completed based solely on internal conditions. The system does not wait passively for external commands.

Distributed

Each element has its own independent decision axis, so that no single point of failure is created.

Asynchronous

The system does not assume global synchronism. Instead, local optima accumulate and stabilize the system as a whole.

These are not emotional notions like “safety” or “peace of mind,” but design principles grounded in control theory and distributed systems theory.

VIII. Why This Architecture Withstands Long-Term Operation

The reasons this architecture can withstand long-term operation are straightforward:

  • Control logic is simple.
  • The number of external variables is small.
  • Failures remain localized.
  • Maintenance and inspection scopes are limited and clear.

The system is designed on the premise of unattended, 10-year, 24/7 non-stop operation. Importantly, this is not just a design concept — it is already proven in real operation.

IX. Relationship with Smart Meters and Smart Grids

In this system, the relationship among the three concepts is organized as follows:

  • Smart meters are observation devices.
  • Smart grids are auxiliary control concepts.
  • Off-grid is the foundational structure.

Between 2010 and 2015, we collected primary data from our smart meters at a scale that was, at the time, too large for the then-available hardware and software to fully process.

However, advances in machine learning and computing resources since around 2018 have turned this data into an extremely valuable asset for optimally designing off-grid systems.

Smart meters and smart grids thus play their role not as ends in themselves, but as supporting technologies that help refine the design of off-grid systems.

X. Personal Energy® as a Product

Starting with the launch of BMS576 in 2011, the system has evolved step by step. The current generation, “JIZAI”, has successfully off-gridded factories, logistics hubs, and even entire small towns.

Our portable UPS products are derived from this architecture and serve as simplified, practical implementations — they are not the main structure itself.

In other words, the products are not the objective. They exist as the result of the design.

XI. Conclusion

Off-grid is not the name of a specific technology. It is the design solution naturally reached when constraints are correctly set and the system is designed accordingly.

The system that remains is:

the structurally simplest and most stable architecture from an engineering perspective — that is off-grid, and that is Personal Energy®: a conversion system that realizes the “role reversal” in which every energy consumer becomes an energy producer.

Off-grid is not a romantic ideal. It is simply the system that survived rigorous engineering constraints.