Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Motor driven rod pumping unit construction.

More »

Fig 1 Expand

Fig 2.

Physical structure of rod pumping unit.

More »

Fig 2 Expand

Fig 3.

Schematic diagram of converting DCs into motor power curve.

More »

Fig 3 Expand

Fig 4.

Automatic adjustment of input data by electrical reference production model.

More »

Fig 4 Expand

Fig 5.

Structural diagram of N-HiTS prediction model.

More »

Fig 5 Expand

Fig 6.

Contrast of N-HiTS and N-BEATS blocks.

More »

Fig 6 Expand

Fig 7.

Local predictions are combined into global predictions.

More »

Fig 7 Expand

Fig 8.

Rod pumping unit.

(A) Structure of rod pumping unit. (B) Drive motor and gearbox.

More »

Fig 8 Expand

Fig 9.

Drive motor and parameter sensor.

(A) Drive motor. (B) Sensors installed in the distribution box.

More »

Fig 9 Expand

Table 1.

Parameters for driving the motor.

More »

Table 1 Expand

Fig 10.

DC and converting the resulting electrical power curve.

(A) Normal. (B) Insufficient fluid supply. (C) Gas influence. (D) Leakage from the pump.

More »

Fig 10 Expand

Fig 11.

Statistical results of measured motor power data in one day.

More »

Fig 11 Expand

Fig 12.

Electric power time series prediction results after sampling.

More »

Fig 12 Expand

Fig 13.

Motor power time series predicted by different forecasting methods.

More »

Fig 13 Expand

Table 2.

Evaluation of prediction results of actual motor power parameters.

More »

Table 2 Expand

Fig 14.

A combination of three lengths of motor power time series.

More »

Fig 14 Expand

Fig 15.

Experimental results with a predicted length of one period.

(A) The motor power sequence length is 8 days. (B) The motor power sequence length is 12 days. (C) The motor power sequence length is 16 days.

More »

Fig 15 Expand

Fig 16.

Comparison of model prediction results.

(A) An 8-day evaluation index of prediction results. (B) An 12-day evaluation index of prediction results. (C) An 16-day evaluation index of prediction results. (D) The N-HiTS model compared with the predictions of other models.

More »

Fig 16 Expand

Table 3.

A three-length sequence of parameters is used to predict the time spent in seconds.

More »

Table 3 Expand

Fig 17.

Multi-period forecast results.

More »

Fig 17 Expand

Fig 18.

Evaluation of Multi-period prediction results.

(A) An evaluation index with a prediction length of 2 days. (B) An evaluation index with a prediction length of 3 days.

More »

Fig 18 Expand

Fig 19.

Prediction of composite fault states.

(A) Prediction of the composite state of gas influence and overbalance. (B) Prediction of the composite state of traveling valve leakage and underbalanced.

More »

Fig 19 Expand

Fig 20.

The ratio of smaller peak to larger peak in the prediction experiment.

More »

Fig 20 Expand