MONITORING OF FUEL FOR POWER LIMITED DEVICES

Related Applications

This application is related to US Provisional Application Serial No.

61223697 filed 07-JUL-2009, as well as US Application Serial No. 12498984 filed 07-JUL-2009 and US Provisional Application 61050620 filed 05-MA Y- 2008. Applicant claims priority of these applications, and the contents of these applications are incorporated herein as if fully recited herein

Field of the Invention

This invention relates to fuel in mobile vehicles, and particularly to fuel monitoring that needs to operate over extended periods even when the vehicle engine is off.

Background of the Invention

The ever increasing cost of gasoline and diesel fuel has created a growing need by operators of equipment using these fuels to monitor their inventories in real time. Absent constant monitoring, fleet owners are easy prey for fuel thieves and dishonest fuel distributors.

Typical motorized equipment contains a rechargeable battery used for providing the energy necessary to operate a starter motor when it is desired to turn the engine on. Mechanically connected to the engine is an alternator which provides the energy to recharge the battery and operate any ancillary loads required by the system. When the motor is not running, power for all loads must come from the battery. Over extended periods of time with the engine off, these loads may siphon off so much energy that there may not be enough capacity left to restart the engine. Ultimately, the battery may suffer irreversible damage due to excessive discharge.

In a mobile asset, fuel sloshing due to vehicle acceleration results in the measured fuel level never being constant. Time-based filters are typically implemented to remove the sloshing artifact so as to give a meaningful measure of the fuel level. For a filter to operate properly, it must have a continuous stream of data on which to operate. This requires that the fuel sensor and any readout electronics must be powered continuously, which may not be practicable given the available energy capacity of the storage battery.

An object of the invention is to avoid these difficulties, and provide fuel monitoring that operates over extended periods even when the engine is not charging the vehicle's battery.

Summary of Embodiments of the Invention

According to an embodiment of the invention, such objects are attained by monitoring the fuel discontinuously while still affording accurate detection of fuel anomalies.

These and other features of the invention are pointed out in the claims. Other objects and advantages of the invention will be evident from the following detailed description when read in light of the accompanying drawings.

Brief Description of the Drawings

Figure 1 is a schematic presentation of a system embodying the invention and involving fleets of vehicles. Figure 2 is a schematic presentation of a system embodying the invention showing details of vehicles in Figure 1.

Figure 3 is a graph illustrating an idealized waveform of fuel level shown for purposes of illustration.

Figure 4 is a graph illustrating discontinuous sampling of a fuel waveform and identifying the nomenclature used for describing embodiments of the invention.

Figure 5 shows a sample fuel loss waveform.

Figure 6 shows the data acquisition flow chart

Figure 7 shows the packet qualification flow chart

Figure 8 shows the packet analysis flow chart

Detailed Description of Preferred Embodiments In Figure 1, a number of customers Customer 1, Customer 2, . .

.Customer m own m fleets Fleet 1, Fleet 2, . . .Fleet m of vehicles, AVl, AV2 . .

.AVn; BVl, BV2, . . . BVn; and CVl, CV2, . . . CVn.

For simplicity any vehicle may be referred to as Vy. Each vehicle Vy contains a transmitter ATx, . . . BTx . . . CTx, where x is a number 1, 2, . . . n. The vehicles communicate with a Data Center DCl, which in turn relays information to the customers Customer 1, Customer 2, Customer n. Each of the vehicles also contains a fuel tank and a sensing system.

Figure 2 illustrates a vehicle Vy with a fuel tank and sensing system in communication with the data center DC and a customer. In Fig. 2 a fuel tank 1 is any of a variety of fuel tanks with known geometry, containing fuel to a level shown as 3. The fuel level is sensed by a fuel sensor 2, which outputs a voltage proportional to the fuel level. The fuel sensor 2 may employ any of a variety of sensor technologies, including but not limited to: float, ultrasonic, capacitive. The voltage is digitized by analog-to-digital converter 4 under control of microcontroller 5. The microcontroller 5 contains a program that processes the samples and determines if fuel theft, fuel gain or expected has occurred.

In the event of an adverse or other reportable event, microcontroller 5 creates a message, which is sent over a wireless link 6 by radio transmitter 9.

This message is received by radio receiver 7 in the data center DCl and the appropriate customer Customer 1, Customer 2, or Customer 3, is notified in any of a number of ways. These may include but are not limited to: website alert, cell phone message, voice message, pager, email. The wireless link 6 may be any of but not limited to: satellite, cell phone, proprietary RF.

According to an embodiment of the invention data center DCl is eliminated, and signals from any vehicle Vy in a customer's fleet travels directly to the customer without passing through data center DCl.

The fuel sensor is powered from a voltage source 10, which typically is the asset's rechargeable battery. Power is switched on or off by a switch 11, which is controlled by the microcontroller 5. The microcontroller operates to open and close the switch 11 between the sensor 3 and the sensor source 11, such as a vehicle battery, according to a program that samples the sensor 3 rather than conducting a constant sensing of the sensor 3. As stated, the microcontroller 5 program processes the samples and determines if fuel theft, fuel gain or expected has occurred. Figure 3, illustrates the waveform of fuel levels sensed by a sensor such as sensor 3. Fuel waveforms are composed of three superimposed signals. The first is a high frequency component (typically 1 Hz) caused by "flutter" on the surface of the fuel. This is typically caused by engine vibrations. The second is a low frequency component (typically 20 second period) caused by bulk movement of the fuel within the tank ("sloshing") due to motion of the vehicle. The graph in Figure 3 is idealized waveform. A third waveform does not appear in Figure 3. This is a slow decrease in fuel level (typically 1 gallon/hour) caused by usage of the fuel by the engine.

According to an embodiment of the invention, the microcontroller 5 closes and opens the switch 11 between the sensor power or battery 10 and the fuel sensor 2 and discontinuously allows the battery 10 to activate the sensor 2.

An a/d converter converts the sensor output into digital format and applies the sensed output to the microcontroller 5 for the period that the microcontroller has closed the switch 11. According to an embodiment, the output of the sensor 2 is digital and the a/d converter is removed from the system. According to an embodiment, the microcontroller 5 itself utilizes the analog data from the sensor

2 and converts it directly, thereby also obviating the need for the a/d converter 4. The opening and closing of the switch 11 by the microcontroller 5 forms bursts or packets during which sampling occurs. Figure 4 illustrates the discontinuous sampling of the fuel waveform and shows the nomenclature of this system.

Waveform analysis is performed on a packet-by-packet basis, with compensation made for the discontinuous basis by which the packets are acquired.

As stated, according to an embodiment of the invention, the system acquires data discontinuously in bursts. This is accomplished by microcontroller 5 closing the switch 11 during each burst to connect the battery or sensor power

10 to the sensor 3 and sample the output of the fuel sensor 3. The sampling rate during the bursts is selected such that the high frequency component of the waveform is sampled in accordance with Nyquist' s theorem (at least twice the highest frequency component present). The sample time is selected so that several cycles of the low frequency component may be acquired. The burst rate is governed by the power budget of the system: when running off the battery 10 , bursts may occur infrequently so as to reduce the power drain from the battery; when the vehicle's engine is running, bursts may occur on a continuous basis because the engine charges the battery and power drain is not an issue.

According to an embodiment, the power requirement of the fuel sensor 3 dominates the power budget of the system. According to an embodiment, for those fuel sensors requiring a relatively large amount of current the burst rate is lower than for those using lesser amounts of current. In one embodiment using an efficient fuel sensor 3, the sample rate is 4 Hz, the sample time is 2 seconds, and the burst rate is 4 seconds. In another embodiment using a less efficient sensor, the burst rate is reduced to 20 seconds.

Algorithm Description: Sampling and processing fuel waveforms from static assets is rarely a problem for fuel monitoring systems. The difficulty arises when the fuel level is dynamic, caused by vibration and motion. The algorithm according to an embodiment of the invention includes filtering to remove such artifact and also detection to decide that the samples within a packet are too unstable to be accurately processed.

The microcontroller 5 is arranged to assume that fuel sloshing has a zero average. It takes sample windows to be long enough so as to allow simple averaging of the samples, thereby filtering out most sloshing artifact and allowing the packet to be used for later analysis. The microcontroller 5 uses peak-to-peak analysis on the samples within a packet to discard those packets whose data is unstable. The microcontroller 5 computes all fuel loss or refuel events from known good packets, even if the determination of an event needs to be delayed.

Figure 5 shows a sample fuel loss waveform as monitored by the microcontroller 5. Packets Pl - P7 are discontinuous sample windows taken of this waveform. Packet Pl is a stable waveform, and would be stored by the microcontroller 5 as the last known good packet, OLDFUEL. Packets P3, P4, P5 show too much variation, and microcontroller 5 would discard them. Packet P6 is once again a good packet, and this would be compared with packet Pl by the microcontroller 5 to declare a fuel loss event. Packet P6 becomes the new OLDFUEL to be used for reference. Packet P7 is the next good packet, and when compared with the current OLDFUEL (P6) is found to be below the loss or gain threshold and therefore a fuel event is not declared. P7 becomes the next OLDFUEL for the next packet.

The fuel sensor is powered from a voltage source 10, which typically is the asset's rechargeable battery. Power may be switched on or off by a switch 11, which is controlled by the microcontroller 5. According to an embodiment of the invention, the switch 11 is a semiconductor switch.

The software operation of the microcontroller 5 may be divided into two sections. The first is data acquisition. The microcontroller 5 acquires data in real time while closing the switch 11 until the microcontroller 5 obtains a complete packet. Data acquisition is shown in Figure 6. The microcontroller 5 performs some simple analysis during this time to identify error conditions resulting in data being blatantly out of range. Packet qualification is shown in Figure 7. The microcontroller 5 performs data packet analysis at the completion of the data acquisition of each packet. The microcontroller 5 compares data from the new packet to that from previous packets to make determinations of fuel events. Packet analysis is shown in Figure 8.

In the Data acquisition flowchart in Figure 6, the microcontroller 5 acquires data at the specified sample rate for the specified sample time to make up a data packet. According to an embodiment, as they are acquired, samples receive some pre-processing to ease the implementation. The preprocessing may occur in the a/d converter or a preprocessor.

The active range of the fuel sensor 2 output is less than the dynamic range of the A/d converter 4. Electrical failures of the sensor 2 or the connections can result in a signal being out of range and erroneous fuel readings. These are trapped and counted in step 20. If too many occur, a message may be sent alerting service personnel of a hardware failure.

The sample counter in step 21 is incremented with each sample to determine when all the data for a given packet is complete.

At the conclusion of a packet's data acquisition, the microprocessor 5 computes the average value. According to an embodiment, the microcontroller 5 stores all the values and computes the arithmetic mean at the end. According to an embodiment, the microcontroller 5 does this more efficiently by accumulating the values in step 22 and then doing the division at the end. According to an embodiment, the microcontroller 5 sets the packet size is set to be 2 ⁿ , allowing a simple shift operation at the end to accomplish the division. FuelMax in step 22 and FuelMin in step 23 are peak and valley detectors. They store the values of the largest and smallest samples and are used in a later stage to qualify a given packet. In step 33 of the packet qualification flowchart of Figure 7, the microcontroller 5 converts sensor 2 output to gallons in the tank. This correction, which may be linear or non-linear corrects for the shape of the fuel tank and linearity of the sensor. In step 30 the microcontroller 5 determines if too many samples for a given packet were out of range, that packet is suspect and discarded.

In step 31 f there was too much variation between samples in a given packet, perhaps due to excessive sloshing, then that packet is discarded.

In step 32 if a packet appears to be valid, the arithmetic mean of the samples is computed.

In the packet analysis flowchart of Figure 8, the average value of the samples in a good packet are compared with the average value of the last known good packet. Variations exceeding positive and negative thresholds are indicative of fuel loss or refuel events.

A series of discarded packets due to motion artifact while fuel is being consumed could give rise to a false fuel theft detection at such time as a good packet is received some hours later. For this reason, a fuel usage estimator is included.

The current value of OLDFUEL is used for reporting fuel level if desired.

For purposes of power budget calculations, assume that the system can sustain an average current draw of 2 ma and that the sample time is 2 minutes. Thus, for a sensor which draws 4 ma, the system would be set up for a sample period of 4 minutes (50% duty cycle, giving an average current of 2 ma)

For another sensor which draws 40 ma, the system would be set up for a sample period of 40 minutes (5% duty cycle, giving an average of current of 2 ma).

The invention overcomes the problems in previous approaches such as described in the aforementioned US Application Serial No. 12498984 filed 07- JUL-2009, which relied on continuous sampling of the fuel level. Kobayashi et al in US Patent 4,470,296 entitle Fuel Gauge For An Automotive Vehicle also use continuous monitoring. In the case of mobile refrigeration equipment (reefers) for instance, the vehicle may travel for extended periods of time when the engine is off. Cerruti US Patent 4,912,646 entitled Method For Detecting The Fuel Level In The Tank Of A Motor Vehicle presents an algorithm based on continuous sampling. Nawrocki in US 5,072,615 entitled Apparatus And Method For Gauging The Amount Of Fuel In A Vehicle Fuel Tank Subject To Tilt disables sampling when the tank is tilted, but still uses continuous sampling. The present invention overcomes these deficiencies, it is not acceptable to wait until the fuel has stopped sloshing before making a level determination.

While embodiments of the invention have been described in detail, it will be recognized that the invention may be embodied otherwise.