Closed-Loop Daylight-Sensing Algorithm for Horizontal Venetian Blinds

Responsive Daylight Control

Responsive daylight control, in which window shading is automatically adjusted to maintain a desired level of glare-free natural illumination, is the key to unlocking the full energy-saving, ergonomic, health, and wellness benefits of daylighting.

There are two key aspects of a responsive daylight control system: the shading device and the control strategy:

• The most cost-effective type of shading device for responsive daylight control is the horizontal blind, because a horizontal blind’s slat-tilt function is exceptionally easy to motorize and provides excellent (and relatively unobtrusive) control of admitted daylight. For this and other reasons, motorized horizontal blinds are poised to dominate the automated shading market.
• The simplest and most effective control strategy for most responsive control applications is closed-loop control, in which the process variable (in this case, the admitted daylight) is sensed directly—via a daylight sensor located in the daylit space—to provide control feedback.

See our dedicated post on responsive daylight control for why it’s potentially the most valuable form of automated shading, but also virtually non-existent in today’s buildings because it’s so hard to implement cost-effectively with conventional technology.

The Problem with Conventional Technology

There are two challenges in closed-loop daylight control with a venetian blind—one of which can be overcome with conventional technology, and one which can’t.

The first challenge is that, by definition, the sensor is located in an area which can receive both artificial and natural illumination. Fortunately, there are at least a few ways to ensure that the sensor output reflects just the daylight level:

• The sensor can be oriented and shielded to minimize the amount of artificial illumination that reaches it.
• If the amount of artificial illumination can be estimated (e.g. via information from the lighting system), it can be subtracted from the sensor output to yield the daylight level.
• The spectral response of the sensor can be chosen to span a portion of the daylight spectrum while excluding visible wavelengths; the sensor will then respond only to daylight while ignoring illumination from LED and fluorescent lamps. We invented this approach (U.S. patent 6,084,231) and have since perfected it in the form of our multi-spectral glare-sensing technology.

Unfortunately, the second challenge is harder to overcome: tilting a blind’s slats doesn’t just change the amount of the admitted daylight—it also changes the direction of the admitted daylight.

This isn’t an issue if the closed-loop sensor is positioned to “see” the blind from the same perspective as the room’s occupants, but unfortunately that’s typically not feasible.  And if the sensor isn’t so positioned, its output could be increasing when the admitted daylight is decreasing, and vice-versa.

For example, the following figure compares the output of two closed-loop daylight sensors as a function of the slat tilt angle of a horizontal blind which is receiving direct sunlight:

Sensor S1 is co-located with a room occupant and “sees” the admitted daylight in the same way the occupant would perceive it.  The output of sensor S1 is therefore well-correlated with the perceived daylight level, facilitating closed-loop control.  Unfortunately, this configuration requires a remote sensor with RF connectivity, increasing cost.  Even more significantly, it’s generally not possible to position such a sensor unobtrusively.

On the other hand, sensor S2 is mounted at the top of the blind and faces inward to sense daylight reflected from surfaces in the room.  This is a far more attractive configuration from an implementation standpoint because the sensor is out of the way.  Even better, such a sensor can even be integrated into a headrail-mounted blind-automation product (like our IntelliBlind™ smart miniblind actuator), thereby eliminating the need for a remote sensor.

However, because of how S2 is positioned and oriented, its output doesn’t always vary in the same way as the daylight level perceived by the occupant.

This is especially true over the range of slat tilt angles that tend to obscure the sky view, which is the optimum range for responsive daylight control because it enables complete blocking of direct sunlight (which is critical to control glare).

The problem is that when the slats are tilted to block the sky view, a sensor positioned as S2 will receive daylight reflected from the tilting slats and from the ground outside the window—neither of which reaches the room occupants.

The result is that the sensor output can often be rising when the perceived daylight level is falling, and vice-versa.  For example, Figure 1 shows the typical output when low-angle direct sunlight is incident on the blind.  Reflections from the slats not only cause S2’s output to rise as the slats are tilted to block the sky view, they also cause ripples in the output-versus-tilt curve.

If the output of S2 were used for closed-loop control under these conditions, the slats would drastically over-close or over-open with changes in the daylight level—a behavior which is highly annoying to users.

It is possible to use a baffle and lens to partially mitigate the effects of spurious daylight components reaching the sensor, but that increases the cost and size of the sensor, hampering product integration.

Our Solution

We discovered the solution to this problem through extensive testing of prototype daylight control systems.  The testing showed that accurate photometric control of daylight is relatively unimportant to users, provided that one key requirement is met: the automatic adjustments must appear to make sense.  Specifically, adjustments that reduce the admitted daylight must be preceded by an increase in the daylight level, while adjustments that increase the admitted daylight must be preceded by a decrease in the daylight level.

This observation, in turn, led to the discovery that effective closed-loop control of a horizontal blind with a sensor output such as the S2 curve of Figure 1 is possible by simply adjusting the sensor output as a function of slat tilt.

Figure 2 shows the approach and the results under the same conditions of Figure 1.  The sensor output is multiplied by an adjustment which is a simple predetermined function of the slat tilt (the tilt does not have to be an absolute tilt angle in degrees, but can instead be a relative tilt setting, e.g. percentage of full tilt).

The adjustment function reduces the effects of the spurious daylight components reaching the sensor as the slats are tilted to block the sky view.  The desired daylight component (the one due to admitted daylight reflected backward from the room surfaces) is also reduced, but the overall effect is that the sensor output still tracks the perceived daylight level well enough for highly effective closed-loop control.

That’s evident in the adjusted S2 output curve of Figure 2.  Note that the curve matches the S1 curve much more closely than the un-adjusted S2 curve of Figure 1.  While the adjusted S2 curve doesn’t perfectly match the S1 curve, the mismatch isn’t enough to result in a noticeable difference in closed-loop daylight control between the two sensors.  Most importantly, there is no extreme over-closing or over-opening of the slats when the adjusted S2 curve is used for closed-loop control.

Further, while the examples of Figures 1 and 2 are for just one sky condition and sensor/blind configuration, the same adjustment function is effective under a wide range of sky conditions and for any sensor mounted at or near the blind (regardless of blind size or room dimensions).  With an appropriately modified adjustment function, the approach is equally effective with sensors mounted a considerable distance away from the blind, such as on the ceiling near the center of a room.

Advantages

Our closed-loop daylight-sensing algorithm is far more effective than baffles or optics at mitigating the effects of spurious daylight components that can reach a closed-loop daylight sensor used with a horizontal blind.  And unlike those prior-art approaches, it’s “free” because it can be implemented purely in software with a just a few lines of code.

Complementary Technologies

While this technology overcomes the unique challenges associated with use of horizontal blinds for closed-loop daylight control, it’s even more advantageous when combined with two of our other innovations aimed at responsive daylight control:

• Our multi-spectral daylight sensor is a highly cost-effective means of reliably sensing daylight glare. It can be used for both open-loop and closed-loop control; the closed-loop configuration is insensitive to artificial illumination and is the perfect complement to our closed-loop daylight sensing algorithm.
• Our fluctuation-mitigation technology enables a responsive daylight-control system to respond quickly to isolated changes in the daylight level while ignoring sustained high-amplitude fluctuations due to moving clouds (which would otherwise cause excessively frequent shading adjustments).

Together, these innovations can unlock the full potential of horizontal blinds for responsive daylight control.

Applications

Our closed-loop daylight-sensing algorithm is advantageous for any smart-blind system aimed at regulating interior daylight levels. Because it facilitates closed-loop control, it enables use of a daylight sensor that’s located on the room-side of the blind, instead of a conventional open-loop sensor which must be located on the window-side of the blind. This not only facilitates installation, it also enables the sensor to be integrated into a headrail-mounted automation device, such as our IntelliBlind™ smart miniblind actuator or IntelliLux™ smart headrail sensor:

With an appropriate choice of adjustment function, the algorithm can also be used with a ceiling-mounted or wall-mounted sensor.

Because the algorithm requires the sensor output and the relative slat tilt setting as inputs, it’s most conveniently implemented in a processor that already has that information (as is the case with the microcontroller in IntelliBlind™ or any smart-blind product with an integrated daylight sensor). When a separate sensor is used (e.g. IntelliLux™), the algorithm can be hosted in the sensor’s microcontroller, the blind-motorization device’s microcontroller, a local automation hub, or in the cloud, with the most convenient implementation depending on the wireless protocols in use.